ProgrammableMarketing/metorial.com (opens in a new tab)
PDF

Views

  • Plan
  • Roadmap
  • Research

On this page

  • The brief
  • Start here
  • Idea portfolio
  • Effort vs impact
  • This week · 7
  • This quarter · 10
  • Big bets · 3
  • Wildcards · 3
Generated Nov 10, 2026
  1. Programmable
  2. Playbooks
  3. Metorial
Metorial logo

growth playbook · Nov 10, 2026

metorial.com (opens in a new tab)

Metorial growth playbook

Diagnosis

Metorial is a four-person YC F25 team with a big product surface: 1,000+ integration pages, 146 scenarios, 45 answer pages, 25 comparison pages, 3.4k GitHub stars and a near-daily changelog. The pitch moved upmarket in the last year ("Put AI to work across your company, safely", starting price $35 to $250).

The funnel has not caught up. The homepage and scale-ups page sell SSO and group-based access, but the only self-serve paid plan leaves both out and Enterprise has no price, so every interested team has to book a demo with four people.

Proof is thin: two customer stories without numbers, and a demo page with no proof at all. This playbook first makes the core promise buyable, adds proof and fixes catalogue consistency, then turns the open-source, CLI and Dev-plan audiences into qualified pipeline. Later bets are a shadow-MCP scanner, a public server directory and a yearly usage report.

Growth Score

6/10

Biggest Opportunity

Make governed company-wide AI access something a scale-up can buy without a sales call, then use the large catalogue and open-source reach to bring those buyers in.

Biggest Risk

Well-funded rivals with logo walls and case studies (Runlayer, Arcade, Composio, Obot) sell the same governance story, and a four-person team that relies on demos for every SSO deal will lose on speed and proof.

01

Start here

The three moves to make first. Each one is written up in full below.

  1. 1

    Put SSO and group access in a published plan: the homepage sells it, but the $250 Scale plan leaves it out

    List the last 20 demo requests and mark how many asked mainly for SSO and group access at fewer than 200 seats.

    High impact First result in 2-4 weeks
  2. 2

    Use one integration count and the right brand names everywhere: 600+, 1,000+ and 1,200+ are all in use

    Update the README and YC launch text to the homepage number.

    Medium impact First result in 1 week
  3. 3

    Run a weekly consistency check on the 26 comparison pages, starting with the MintMCP page that contradicts itself

    Correct the MintMCP FAQ answer to match the table.

    Medium impact First result in 1 week

02

Your idea portfolio

23 ideas, grouped by when to run them. Open any idea for how to run it: the steps, who to approach, copy you can use, and how to measure it.

  • 8 ideas are marked Unexpected: plays you are unlikely to have tried.

Every idea placed by the effort it takes against the impact it could have. Quick wins sit top left; select a number to open that idea. The filters below narrow the list by status, timing, category and what Programmable can build.

Effort vs impactEach number is an idea. Select one to open it.
Impact →
Quick wins · 1Big bets · 9Fill-ins · 10Reconsider · 3
Effort →

Showing 23 of 23 ideas

This week 7

Quick wins with a clear first action.

  • 01

    Metorial's search snippet opens with "One MCP URL behind your SSO" and "role-based tool access", and the scale-ups page promises "company SSO" with policies "per user and group". On the pricing page, Scale ($250/mo, "Businesses and teams that need more") marks SAML SSO and access management as not included. Both are Enterprise only, which has no price, and both the scale-ups and SAML pages end in "Book a demo". A 60-person scale-up that wants exactly what the homepage describes cannot buy it, and has to wait for a four-person team to run a sales call. Rivals take that buyer: Obot gives away a self-hosted edition for up to 100 users.

    High impactMedium effort First result in 2-4 weeks Founder/CEO

    First action

    List the last 20 demo requests and mark how many asked mainly for SSO and group access at fewer than 200 seats.

    How to measure it

    Self-serve SSO plan upgrades

    Now
    –
    Target
    10 new paying teams
    By
    90 days

    Steps

    1. 1

      Size the demand

      Tag recent demos and inbound emails by what they wanted (SSO and groups, on-prem, SLA, volume) and by seat count.

      Founder/CEO 1 day

    2. 2

      Choose the shape

      A middle tier or an add-on to Scale; set a price per month or per seat using the demo answers and rivals' published plans.

      Founder/CEO 2 days

    3. 3

      Check the product gates

      Confirm SAML, SCIM groups and policies can be switched on per plan in the platform, with no manual setup.

      Founder/CTO 3 days

    4. 4

      Add the SSO trial

      14 days of SSO and group policies on Scale, with an in-app countdown and an upgrade button.

      Founding Engineer 1 week

    5. 5

      Rewrite the pricing table

      A new column or add-on row, "from $X" on Enterprise, and a short FAQ on what needs Enterprise.

      Founder/CEO 1 day

    6. 6

      Repoint the SSO pages

      The scale-ups, SAML and access-control pages and the homepage FAQ link to the plan, with "Book a demo" as the second button.

      Founder/CEO 1 day

    7. 7

      Email open demo leads

      Tell recent mid-sized leads they can now start with SSO the same day.

      Founder/CEO 1 day

    8. 8

      Measure

      Pricing-page upgrades, trial-to-paid, and demos per closed deal, before and after.

      Founder/CTO monthly

    First result
    2-4 weeks
    Cost
    About two weeks of founder and engineer time; no extra spend
    Owner
    Founder/CEO

    Smallest test

    Add an "SSO add-on: $X/mo, start today" row to Scale for two weeks, fulfilled by hand, and count clicks and payments.

    Why it works

    Publish a plan that includes the homepage promise. One option is a "Team" or "Business" tier between Scale and Enterprise with SAML SSO, SCIM group sync, group policies and audit-log export at a listed per-month price. The other is an SSO and access add-on to Scale. Keep Enterprise for on-prem, SLA, advanced compliance and custom limits, and show "from $X" so buyers can budget. Add a 14-day SSO trial on Scale so IT can connect its identity provider before paying. Then point the scale-ups page, the SAML and access-control feature pages and the homepage FAQ to the new plan instead of only "Book a demo". Track self-serve upgrades and how many demos turn into this plan.

    The pricing page is where intent peaks. A buyer who reads "behind your SSO" on the homepage and finds "No" in the SSO row on Scale either leaves or waits for a call. Publishing the price turns the most common mid-market request into a self-serve purchase, shortens sales cycles for a team with no sales staff, and keeps demos for real enterprise deals. It also matches the move upmarket the site already made (starting price $35 to $250 in a year).

    Ready to use

    CopyDraft pricing row

    Feature | Dev | Scale | Team (new) | Enterprise
    Price | $0 | $250/mo | $X/mo or $Y/seat | from $Z
    SAML SSO | No | 14-day trial | Yes | Yes
    SCIM group sync | No | No | Yes | Yes
    Group policies and tool-level access | No | No | Yes | Yes
    Audit-log export | No | No | Yes | Yes + SIEM
    On-prem, SLA, advanced compliance | No | No | No | Yes
    How to start | Sign up | Sign up | Sign up | Book a demo

    Risks and how to handle them

    • Enterprise deals downgrade to the cheaper plan

      Cap seats and keep on-prem, SLA, advanced compliance and custom retention in Enterprise.

    • Support load from self-serve SSO setups

      Ship a setup guide per identity provider and a test-connection button before launch.

    • Price set too low

      Publish it as a launch price and review it after 20 sign-ups.

    Evidence

    • Your websitePricing page: Dev "$0/mo" for "Small projects and personal use" (500,000 tool calls, 2 team members, no production access); Scale "$250/mo" for "Businesses and teams that need more" (2,500,000 tool calls, 20 team members, production access, "basic compliance", support via Discord and email), with SAML SSO and access management marked not included and overage of "$1 per 1,500 additional tool calls"; Enterprise is custom-priced and is the only plan with "SAML, SSO and IAM", "Granular permissions (RBAC)" and "Advanced compliance (SOC 2 II, GDPR)". No free trial, annual option or FAQ; developers are told to "npm install -g @metorial/cli".Source
    • Your websiteScale-ups page: Metorial "helps scale-ups roll out AI agents with company SSO and one Magic MCP link", with policies "per user and group"; it names no plan or price and its only call to action is "Book a demo". The SAML SSO feature page also names no plan or price ("Book a demo, or talk to sales").Source
    • Your websiteHomepage (checked November 9, 2026): H1 "Put AI to work across your company, safely"; meta description "One MCP URL behind your SSO. 1,000+ integrations, role-based tool access, audit logs and shared skills, in Claude, ChatGPT and Cursor. Open source."; one named customer quote (Moonfire); the FAQ states SOC 2 Type II and GDPR compliance with EU and US regions, and that Metorial can be self-hosted or run on Metorial Cloud.Source
    • ResearchSite history (compared with November 11, 2025): the starting price went from "$35/mo" to "$250/mo" and the headline changed from "Metorial" to "Put AI to work across your company, safely".Source
    • Competitor researchRivals' own homepages: Runlayer, "The operating layer for AI", shows eight customer logos, four case studies and quotes from named security leaders, with no public pricing; Composio offers "1,500+ apps" and a free plan with 100,000 tool calls a month; Arcade claims "8,000+ permission-aware tools" and "Free to start. Priced by usage."; Obot is open source, with a free self-hosted Community edition for up to 100 users and a contact-sales Enterprise tier.Source
    • HiringY Combinator jobs (open now): Founding Engineer ($100K-$180K, 1-2% equity), Backend Engineer and Infrastructure/DevOps Engineer, all remote; the company is YC F25, founded 2025, a team of 4, based in Vienna. The posts ask for "deep distributed systems knowledge", TypeScript and Go, "Good understanding of AI agents (ideally MCP and tool calls)" and "AWS + HashiStack". The careers page sits on meteor.computer under the name "Meteor Computing".Source
  • 04

    The homepage says "1,000+ integrations", the GitHub README says "1200+", and the YC launch post still says "600+". Generated titles also get brand names wrong ("Highlevel" rather than HighLevel). A security buyer comparing vendors notices when the basic numbers don't agree.

    Medium impactLow effort First result in 1 week Founder/CEO

    First action

    Update the README and YC launch text to the homepage number.

    How to measure it

    Consistent count across owned profiles

    Now
    –
    Target
    100%
    By
    30 days

    Steps

    1. 1

      Pick the number

      Live catalogue count, rounded down.

      Founder/CTO 1 hour

    2. 2

      Update owned profiles

      README, YC, LinkedIn, X, docs.

      Founder/CEO 2 hours

    3. 3

      Feed it from data

      Homepage and comparison tables read the same value.

      Founding Engineer 1 day

    4. 4

      Add display names

      Official brand spelling per integration.

      Founding Engineer 1 day

    5. 5

      Check top titles

      100 most-viewed integration pages.

      Founding Engineer 1 day

    6. 6

      Recheck

      Each quarter.

      Founder/CEO quarterly

    First result
    1 week
    Cost
    $0
    Owner
    Founder/CEO

    Smallest test

    Fix the README and the HighLevel title today.

    Why it works

    Pick one source of truth (the live catalogue count), show it on the homepage, README, YC profile, comparison data and docs from the same number, and update it on a schedule. Add a display-name field to the catalogue so titles use each vendor's official spelling, and check the 100 most-viewed integration titles.

    Consistent facts build trust in a trust-led category, and AI assistants and comparison sites repeat whichever number they find first.

    Risks and how to handle them

    • A rounded number looks smaller

      Show the exact live count on the integrations page.

    Evidence

    • GitHubmetorial/metorial: 3.4k stars, 371 forks, Apache 2.0, 2,927 commits; the README says "Connect any AI model to 1200+ integrations", that the platform is open source and self-hostable, and that the server index holds "more than 5000 MCP servers". metorial/mcp-index: 281 stars, 41 forks, MIT. The homepage says "1,000+ integrations" and the YC launch post "600+ integrations".Source
    • Your websiteIntegration pages: /integrations/highlevel lists 37 tools with inline setup steps for five clients, but /integrations/highlevel/claude returns a 404, and the title spells the brand "Highlevel" while the changelog writes "HighLevel"; /integrations/salesforce lists 14 tools, links to separate client guides, 15 related scenarios and a "Deploy Salesforce" link into the platform, and carries no customer proof. Titles follow "... MCP Server for Claude & Cursor | Metorial".Source
  • 05

    Metorial's comparisons are unusually transparent: sourced facts, a public competitors.json, a "Tell us" link. But the MintMCP page says in its decision guide that the rival's pages "don't state a self-hosted option", while its FAQ and FAQ schema answer "Yes, Mintive documents a self-hosted option." Search engines and AI assistants quote FAQ schema directly.

    Medium impactLow effort First result in 1 week Founder/CEO

    First action

    Correct the MintMCP FAQ answer to match the table.

    How to measure it

    Open inconsistencies

    Now
    –
    Target
    0
    By
    30 days

    Steps

    1. 1

      Fix the known error

      MintMCP FAQ and schema.

      Founder/CEO 1 hour

    2. 2

      Audit all pages

      Table against guide against FAQ.

      Founder/CEO 1 day

    3. 3

      Single source

      Render FAQ answers from competitors.json fields.

      Founding Engineer 2 days

    4. 4

      Add the check

      Weekly job flags disagreements.

      Founding Engineer 1 day

    5. 5

      Show dates

      Per-row "checked" date.

      Founding Engineer 1 day

    6. 6

      Publish corrections

      A short log on the hub.

      Founder/CEO ongoing

    First result
    1 week
    Cost
    $0
    Owner
    Founder/CEO

    Smallest test

    Fix the MintMCP page and audit five more this week.

    Why it works

    Fix the MintMCP page now. Then generate every table, decision guide and FAQ answer from the same field in competitors.json, so they cannot disagree, and add a weekly check that flags any page where they do. Show the "checked" date per row and record each correction in a public changelog.

    Comparison pages are high intent, and their credibility is the product. A visible contradiction gives a rival an easy reason to call the page unreliable.

    Risks and how to handle them

    • Rivals dispute facts

      Link each claim to the rival's own page and fix fast through "Tell us".

    Evidence

    • Your websiteComparisons: a hub of 26 comparisons in six categories, "Checked October 17, 2026", backed by /data/competitors.json and a "Tell us" link for corrections. On the MintMCP page the decision guide says "Mintive's own pages don't state a self-hosted option" while the FAQ, and its FAQ schema, answers "Yes, Mintive documents a self-hosted option."Source
  • 06

    The demo page has no logos, quotes, certifications or response time, and its onboarding steps are labelled "1+", "50+" and "500+" with no explanation. The homepage already claims SOC 2 Type II and GDPR, and there is a customer quote to reuse.

    Medium impactLow effort First result in 2 weeks Founder/CEO

    First action

    Move the Moonfire quote and compliance badges onto the demo page.

    How to measure it

    Demo conversion

    Now
    –
    Target
    +20%
    By
    60 days

    Steps

    1. 1

      Add proof

      Two quotes and compliance badges.

      Founding Engineer 2 hours

    2. 2

      Explain the call

      Agenda and length.

      Founder/CEO 1 hour

    3. 3

      Promise a reply time

      For example, within one working day, EU and US hours.

      Founder/CEO 1 hour

    4. 4

      Fix the step labels

      Say what 1+, 50+ and 500+ mean, or remove them.

      Founding Engineer 1 hour

    5. 5

      Add the free path

      "Start free" for developers.

      Founding Engineer 1 hour

    6. 6

      Measure

      PostHog funnel from page to booking.

      Founder/CTO monthly

    First result
    2 weeks
    Cost
    $0
    Owner
    Founder/CEO

    Smallest test

    Add the quote and badges today and compare two weeks.

    Why it works

    Add the Moonfire and Crunched quotes, the SOC 2 Type II and GDPR badges with a trust-center link, a clear "what happens on the call" (30 minutes, your identity provider and three tools connected live), a reply-time promise, and an "or start free now" link. Replace the unexplained numbers with labelled steps.

    Demo pages convert when they lower risk and set expectations, and the proof already exists elsewhere on the site.

    Risks and how to handle them

    • Reply promise missed

      Set an alert for unanswered bookings.

    Evidence

    • Your websiteDemo page: no customer logos, testimonials, certifications or stated response time; onboarding steps labelled "1+", "50+" and "500+" with no explanation; the "Book a demo" button jumps to an on-page anchor.Source
    • Your websiteCustomers page: two stories, Crunched (engineering software: "Metorial lets us turn customer connector requests into something we can actually ship.") and Moonfire (wealth management: "a single pane of glass across all our data sources"), both dated March 8, 2026, with no metrics and no logo wall.Source
    • Your websiteHomepage (checked November 9, 2026): H1 "Put AI to work across your company, safely"; meta description "One MCP URL behind your SSO. 1,000+ integrations, role-based tool access, audit logs and shared skills, in Claude, ChatGPT and Cursor. Open source."; one named customer quote (Moonfire); the FAQ states SOC 2 Type II and GDPR compliance with EU and US regions, and that Metorial can be self-hosted or run on Metorial Cloud.Source
  • 07

    Scale includes 2.5 million tool calls, with overage at "$1 per 1,500 additional tool calls" and "$1 per 750 additional Metorial Callbacks". Buyers think in people and agents, not tool calls, and the blog already argues for "cost per completed task".

    Medium impactLow effort First result in 3 weeks Founding Engineer

    First action

    Pull median tool calls per active user per day from the platform.

    How to measure it

    Pricing-to-sign-up conversion

    Now
    –
    Target
    +15%
    By
    60 days

    Steps

    1. 1

      Get the medians

      Calls per user per day by team type.

      Founder/CTO 1 day

    2. 2

      Write the formula

      Include overage and plan limits.

      Founder/CEO 1 day

    3. 3

      Build it

      On /pricing.

      Founding Engineer 3 days

    4. 4

      Link to Usage

      "See your real numbers" after sign-up.

      Founding Engineer 1 day

    5. 5

      Pass to sign-up

      Prefill plan choice.

      Founding Engineer 1 day

    6. 6

      Measure

      Calculator use and conversion.

      Founder/CTO monthly

    First result
    3 weeks
    Cost
    Engineering time only
    Owner
    Founding Engineer

    Smallest test

    Add a static table for 10, 50 and 200 people.

    Why it works

    Add a small calculator: people, agents per person and tasks a day in; estimated tool calls, the right plan and monthly cost out, with a link to the Usage dashboard that shows real numbers after sign-up. Use median calls per session from real usage as the default.

    Usage pricing scares budget owners when they cannot predict the bill. One estimate removes the objection and qualifies the plan before sign-up.

    Risks and how to handle them

    • Estimates too low

      Show a range and the overage rate.

    Evidence

    • Your websitePricing page: Dev "$0/mo" for "Small projects and personal use" (500,000 tool calls, 2 team members, no production access); Scale "$250/mo" for "Businesses and teams that need more" (2,500,000 tool calls, 20 team members, production access, "basic compliance", support via Discord and email), with SAML SSO and access management marked not included and overage of "$1 per 1,500 additional tool calls"; Enterprise is custom-priced and is the only plan with "SAML, SSO and IAM", "Granular permissions (RBAC)" and "Advanced compliance (SOC 2 II, GDPR)". No free trial, annual option or FAQ; developers are told to "npm install -g @metorial/cli".Source
    • Your websiteBlog: 38 posts, recently about one a week, with slugs such as "ai-agent-policy-on-one-page" (November 5, 2026), "whose-permissions-does-an-agent-use", "cost-per-completed-task" and "who-owns-an-agent-after-pilot".Source
    • Your websiteChangelog: ten entries between October 18 and November 9, 2026, including "Share skills over Magic MCP", "Usage: see what your agents are doing at a glance", "Send admin setup links for Google and Microsoft integrations" and "Create a Metorial organization from the CLI"; 17 updates in 90 days and no newsletter signup found.Source
  • 08

    Some of Metorial's most useful buyer content (how to connect Salesforce securely, SSO and SCIM for AI tools, an AI agent security questionnaire, rolling agents out to employees) sits under /for-ai-crawlers/. The Salesforce guide links to the integration page, but the integration and feature pages don't clearly link back, so people may never find these guides.

    Medium impactLow effort First result in 2 weeks Founding Engineer

    First action

    Add the Salesforce guide link to /integrations/salesforce and the SSO guide to /features/saml.

    How to measure it

    Internal visits to guides

    Now
    –
    Target
    3x
    By
    60 days

    Steps

    1. 1

      Map the links

      Each answer to its integration or feature page.

      Founding Engineer 1 day

    2. 2

      Check reachability

      Which answers have no internal links today.

      Founding Engineer 2 hours

    3. 3

      Add guide blocks

      On integration and feature templates.

      Founding Engineer 1 day

    4. 4

      Surface the best

      Security, SSO and rollout guides in navigation.

      Founder/CEO 1 hour

    5. 5

      Give an index

      A readable guides page.

      Founding Engineer 1 day

    6. 6

      Measure

      Internal clicks and sign-ups from guides.

      Founder/CTO monthly

    First result
    2 weeks
    Cost
    $0
    Owner
    Founding Engineer

    Smallest test

    Link five guides from their integration pages and compare clicks.

    Why it works

    Check which answer pages are linked from the site. Then add a "Guides" block to each integration and feature page linking to the matching answers, put the strongest ones (security questionnaire, SSO and SCIM, rollout) in the main navigation or docs, and give them a human-friendly path or a clear index.

    The content is already written and targets real buying questions. Internal links send both readers and search traffic to it.

    Risks and how to handle them

    • Moving URLs loses existing traffic

      Keep the URLs and only add links, or redirect one to one.

    Evidence

    • Your websiteAnswer page "How to Connect Salesforce to Claude Securely" (updated August 25, 2026): a five-step CLI setup, a deployment table, five FAQs and cited sources, linking to the Salesforce integration page, Magic MCP, Protoguard and pricing. It sits under /for-ai-crawlers/, one of 45 such answers (for example "ai-agent-security-questionnaire", "sso-and-scim-for-ai-tools" and "roll-out-ai-agents-to-employees").Source
    • Your websiteSitemap (checked November 9, 2026): 128 core pages (including industry pages such as /for/wealth-management and /for/saas, and 23 agent-client pages under /agents/), 38 blog posts, 25 comparison pages, 45 answer pages under /for-ai-crawlers/, 146 scenario pages, 1,000+ integration pages, and 250 integration-by-client setup pages (50 integrations times 5 clients). No pages in languages other than English.Source
  • 09

    Metorial has industry pages such as /for/wealth-management and /for/saas, and its two customer stories come from exactly those kinds of companies: Moonfire in wealth management and Crunched in engineering software. Regulated-industry buyers look for a peer before anything else.

    Medium impactLow effort First result in 2 weeks Founder/CEO

    First action

    Add the Moonfire block to /for/wealth-management.

    How to measure it

    Industry-page demos

    Now
    –
    Target
    +20%
    By
    60 days

    Steps

    1. 1

      Match stories

      Story to industry page.

      Founder/CEO 1 hour

    2. 2

      Write blocks

      Quote, integrations used, link.

      Founder/CEO 2 hours

    3. 3

      Place them

      Above the call to action.

      Founding Engineer 2 hours

    4. 4

      Flag gaps

      Industry pages with no story.

      Founder/CEO 1 hour

    5. 5

      Fill gaps

      From the idea 3 interviews.

      Founder/CEO 1 quarter

    6. 6

      Measure

      Demo rate per industry page.

      Founder/CTO monthly

    First result
    2 weeks
    Cost
    $0
    Owner
    Founder/CEO

    Smallest test

    Add the Moonfire block to one page today.

    Why it works

    Add the Moonfire quote and story link to the wealth-management and finance pages, and the Crunched story to the SaaS page, with the specific integrations each used. As new stories arrive (idea 3), give every industry page at least one.

    Peer proof on the page a buyer actually lands on beats a general customers page they never open.

    Risks and how to handle them

    • Only two stories

      Use an anonymised "a wealth manager" quote until more are approved.

    Evidence

    • Your websiteCustomers page: two stories, Crunched (engineering software: "Metorial lets us turn customer connector requests into something we can actually ship.") and Moonfire (wealth management: "a single pane of glass across all our data sources"), both dated March 8, 2026, with no metrics and no logo wall.Source
    • Your websiteSitemap (checked November 9, 2026): 128 core pages (including industry pages such as /for/wealth-management and /for/saas, and 23 agent-client pages under /agents/), 38 blog posts, 25 comparison pages, 45 answer pages under /for-ai-crawlers/, 146 scenario pages, 1,000+ integration pages, and 250 integration-by-client setup pages (50 integrations times 5 clients). No pages in languages other than English.Source

This quarter 10

The core plays for the next three months.

  • 02

    Metorial has 1,000+ integration pages and 23 agent-client pages, but only 250 integration-by-client setup pages: 50 integrations times 5 clients. For the rest, the steps sit inline, so a search like "connect HighLevel to Claude" has no page to land on (/integrations/highlevel/claude returns a 404).

    High impactMedium effort First result in 6-8 weeks Founding Engineer

    First action

    Rank integrations by platform usage and pick the next 100 to get client pages.

    How to measure it

    Sign-ups from setup pages

    Now
    –
    Target
    +50%
    By
    2 quarters

    Steps

    1. 1

      Rank demand

      Platform usage and search volume for each integration.

      Founder/CEO 2 days

    2. 2

      Check the template

      Make sure each page has unique tools, prompts and steps, not just a swapped name.

      Founding Engineer 2 days

    3. 3

      Generate batch one

      100 integrations times 5 clients.

      Founding Engineer 1 week

    4. 4

      Add clients

      Extend to clients that already have /agents pages.

      Founding Engineer 2 weeks

    5. 5

      Update sitemaps

      New URLs in integration-clients.xml, with internal links from each integration page.

      Founding Engineer 1 day

    6. 6

      Measure and expand

      Indexing and sign-ups per batch, then the next batch.

      Founder/CEO monthly

    First result
    6-8 weeks
    Cost
    Engineering time only
    Owner
    Founding Engineer

    Smallest test

    Publish client pages for 20 mid-tail integrations and compare impressions after four weeks.

    Why it works

    Generate the client setup page from the data already on each integration page (tools, auth method, inline steps) for every integration with real demand, and add clients beyond the first five where the /agents pages already exist (VS Code, Windsurf, Zed, Copilot and others). Each page gets a unique title, the integration's own tools and example prompts, and a "Deploy" link into the platform. Roll it out in batches ranked by search demand and catalogue usage, and add the new URLs to the integration-clients sitemap.

    "Connect X to Y" searches have high intent and little competition for long-tail tools. The data and template already exist, so each page costs almost nothing, and every page ends in a sign-up.

    Risks and how to handle them

    • Thin, near-duplicate pages

      Only publish pages with integration-specific steps and tools; noindex the rest.

    • Client setup flows change

      Render steps from one shared client definition so one edit updates every page.

    Evidence

    • Your websiteSitemap (checked November 9, 2026): 128 core pages (including industry pages such as /for/wealth-management and /for/saas, and 23 agent-client pages under /agents/), 38 blog posts, 25 comparison pages, 45 answer pages under /for-ai-crawlers/, 146 scenario pages, 1,000+ integration pages, and 250 integration-by-client setup pages (50 integrations times 5 clients). No pages in languages other than English.Source
    • Your websiteIntegration pages: /integrations/highlevel lists 37 tools with inline setup steps for five clients, but /integrations/highlevel/claude returns a 404, and the title spells the brand "Highlevel" while the changelog writes "HighLevel"; /integrations/salesforce lists 14 tools, links to separate client guides, 15 related scenarios and a "Deploy Salesforce" link into the platform, and carries no customer proof. Titles follow "... MCP Server for Claude & Cursor | Metorial".Source
  • 03

    Metorial has two customer stories, both from March 2026 and neither with a number. A trade publication wrote that its evidence of demand is "early and largely company-reported". Runlayer, selling the same governance story, shows eight logos and four case studies.

    High impactMedium effort First result in 6 weeks Founder/CEO

    First action

    Pull the ten organisations with the most tool calls last month and ask three for a 30-minute interview.

    How to measure it

    Quantified stories live

    Now
    –
    Target
    3
    By
    60 days

    Steps

    1. 1

      Pick accounts

      Top usage, varied industries.

      Founder/CEO 1 day

    2. 2

      Pull the data

      Usage numbers from the platform to bring to the interview.

      Founder/CTO 2 days

    3. 3

      Interview

      Before and after, plus a quote.

      Founder/CEO 2 weeks

    4. 4

      Write and approve

      One page each, signed off by the customer.

      Founder/CEO 2 weeks

    5. 5

      Place the proof

      Homepage, pricing, demo and the related integration pages.

      Founding Engineer 2 days

    6. 6

      Refresh the two existing stories

      Add numbers to the Crunched and Moonfire stories.

      Founder/CEO 2 weeks

    First result
    6 weeks
    Cost
    Founder time; optional thank-you credits
    Owner
    Founder/CEO

    Smallest test

    Add one measured line to the Moonfire story this month.

    Why it works

    Interview the most active customers and publish three short stories with measured results: integrations rolled out, people connected, keys removed from config files, time to roll out, tool calls governed. Add a logo strip (with permission) to the homepage, pricing and demo pages, and a one-line result on the matching integration pages.

    Security and IT buyers copy what peers have done. Numbers make the story credible and give sales a reference for every demo.

    Risks and how to handle them

    • Customers will not be named

      Offer anonymised stories with industry and size.

    Evidence

    • Your websiteCustomers page: two stories, Crunched (engineering software: "Metorial lets us turn customer connector requests into something we can actually ship.") and Moonfire (wealth management: "a single pane of glass across all our data sources"), both dated March 8, 2026, with no metrics and no logo wall.Source
    • PressRuntimeWire startup spotlight (October 2026): Metorial's "public evidence of demand remains early and largely company-reported"; the company is "betting that as companies give agents access to real systems, identity, permissions and audit trails become shared infrastructure."Source
    • Competitor researchRivals' own homepages: Runlayer, "The operating layer for AI", shows eight customer logos, four case studies and quotes from named security leaders, with no public pricing; Composio offers "1,500+ apps" and a free plan with 100,000 tool calls a month; Arcade claims "8,000+ permission-aware tools" and "Free to start. Priced by usage."; Obot is open source, with a free self-hosted Community edition for up to 100 users and a contact-sales Enterprise tier.Source
  • 10

    Dev is free with 500,000 tool calls but has no production environment and 2 team members. Scale's main unlock is production access and 20 seats. The upgrade moment is predictable (a third teammate, or a first production deploy), but nothing on the site shows the product meeting the user there.

    High impactMedium effort First result in 1 month Founder/CTO

    First action

    Log how many Dev organisations tried to invite a third member or create a production environment last month.

    How to measure it

    Monthly Dev-to-Scale upgrades

    Now
    –
    Target
    +40%
    By
    2 quarters

    Steps

    1. 1

      Instrument limits

      Events for seat, environment and call limits.

      Backend Engineer 3 days

    2. 2

      Design the checklist

      Four steps to production.

      Founder/CTO 2 days

    3. 3

      Build in-app prompts

      At each limit.

      Founding Engineer 1 week

    4. 4

      Add the CLI hint

      When a deploy targets production.

      Founding Engineer 2 days

    5. 5

      Send the usage email

      Owner gets their numbers and the upgrade link.

      Backend Engineer 3 days

    6. 6

      Measure

      Upgrade rate by trigger.

      Founder/CTO monthly

    First result
    1 month
    Cost
    About two engineer-weeks
    Owner
    Founder/CTO

    Smallest test

    Email Dev organisations that hit the seat limit last month.

    Why it works

    Add a "Go to production" checklist in the platform and CLI: create a production environment, invite the team, set policies, check logs. When a Dev user hits a limit, show what Scale unlocks for their project with one-click upgrade, and email the account owner with their own usage numbers.

    Upgrades convert best when they solve the problem the user has right now, in the tool they are using.

    Risks and how to handle them

    • Prompts feel pushy to developers

      Show them only at real limits, once per limit.

    Evidence

    • Your websitePricing page: Dev "$0/mo" for "Small projects and personal use" (500,000 tool calls, 2 team members, no production access); Scale "$250/mo" for "Businesses and teams that need more" (2,500,000 tool calls, 20 team members, production access, "basic compliance", support via Discord and email), with SAML SSO and access management marked not included and overage of "$1 per 1,500 additional tool calls"; Enterprise is custom-priced and is the only plan with "SAML, SSO and IAM", "Granular permissions (RBAC)" and "Advanced compliance (SOC 2 II, GDPR)". No free trial, annual option or FAQ; developers are told to "npm install -g @metorial/cli".Source
    • Your websiteChangelog: ten entries between October 18 and November 9, 2026, including "Share skills over Magic MCP", "Usage: see what your agents are doing at a glance", "Send admin setup links for Google and Microsoft integrations" and "Create a Metorial organization from the CLI"; 17 updates in 90 days and no newsletter signup found.Source
  • 11

    With four people and only engineering roles open, Metorial has no sales team. Its free Dev plan and SDKs (about 9,600 weekly npm downloads) bring in many developers, and a few of them are scouting for a company-wide rollout. That shows as a company email domain, an identity-provider connection attempt, several integrations, or a third invite.

    High impactLow effort First result in 3 weeks Founder/CEO

    First action

    List last month's Dev organisations with a company domain and three or more integrations.

    How to measure it

    PQL-to-paid conversion

    Now
    –
    Target
    15%
    By
    2 quarters

    Steps

    1. 1

      Pick signals

      Domain, invites, integrations, identity-provider attempts.

      Founder/CTO 1 day

    2. 2

      Build the score

      A daily query.

      Backend Engineer 3 days

    3. 3

      Alert the founders

      Top five each morning.

      Backend Engineer 1 day

    4. 4

      Write templates

      Short, personal, usage-specific.

      Founder/CEO 1 day

    5. 5

      Reach out

      Same day.

      Founder/CEO daily

    6. 6

      Tune

      Compare score against conversions.

      Founder/CTO monthly

    First result
    3 weeks
    Cost
    Founder time
    Owner
    Founder/CEO

    Smallest test

    Email ten such organisations by hand this week.

    Why it works

    Define a short product-qualified lead score from platform events, send a daily alert of the top five to the founders, and reply with a personal note that offers the SSO plan (idea 1) or an Enterprise call. Record the outcome to tune the score.

    Founder outreach to warm, active users is the highest-yield sales motion for a tiny team, and the signals already exist in the product.

    Risks and how to handle them

    • Developers dislike sales emails

      Offer help, not a demo, and stop after one reply.

    Builds on

    Evidence

    • HiringY Combinator jobs (open now): Founding Engineer ($100K-$180K, 1-2% equity), Backend Engineer and Infrastructure/DevOps Engineer, all remote; the company is YC F25, founded 2025, a team of 4, based in Vienna. The posts ask for "deep distributed systems knowledge", TypeScript and Go, "Good understanding of AI agents (ideally MCP and tool calls)" and "AWS + HashiStack". The careers page sits on meteor.computer under the name "Meteor Computing".Source
    • ResearchPackage registries: 29 npm packages with about 9,600 weekly downloads, mostly SDK and utility packages, while @metorial/cli gets 11 a week; the Python package "metorial" is at version 2.7.0, released November 9, 2026.Source
    • Your websitePricing page: Dev "$0/mo" for "Small projects and personal use" (500,000 tool calls, 2 team members, no production access); Scale "$250/mo" for "Businesses and teams that need more" (2,500,000 tool calls, 20 team members, production access, "basic compliance", support via Discord and email), with SAML SSO and access management marked not included and overage of "$1 per 1,500 additional tool calls"; Enterprise is custom-priced and is the only plan with "SAML, SSO and IAM", "Granular permissions (RBAC)" and "Advanced compliance (SOC 2 II, GDPR)". No free trial, annual option or FAQ; developers are told to "npm install -g @metorial/cli".Source
  • 12

    The open-source repo has 3.4k stars and 371 forks, and the homepage says Metorial can be self-hosted. Every fork is a team that chose Metorial but may never talk to it. Obot's free self-hosted edition for up to 100 users shows rivals compete for the same teams.

    High impactMedium effort First result in 1 quarter Founder/CEO

    First action

    Add an opt-in deployment registration to the next release, with release notes by email as the reward.

    How to measure it

    Paid self-hosted licences

    Now
    –
    Target
    5
    By
    2 quarters

    Steps

    1. 1

      Define the licence

      What open source keeps and what is paid.

      Founder/CEO 1 week

    2. 2

      Add registration

      Opt-in, privacy-friendly.

      Infrastructure/DevOps Engineer 1 week

    3. 3

      Security channel

      Email for registered deployments.

      Infrastructure/DevOps Engineer 3 days

    4. 4

      Licence keys

      Unlock paid features offline.

      Backend Engineer 2 weeks

    5. 5

      Publish the offer

      README, docs and pricing.

      Founder/CEO 2 days

    6. 6

      Reach out

      Founders email registered deployments.

      Founder/CEO ongoing

    First result
    1 quarter
    Cost
    About a month of engineering
    Owner
    Founder/CEO

    Smallest test

    Add a "Running Metorial yourself? Get support" link in the README and count replies.

    Why it works

    Offer a "Self-hosted Business" licence: enterprise features (SSO, RBAC, audit export), signed releases, upgrade help and an SLA, priced per deployment. Add an opt-in "register this deployment" step that unlocks a free security-update channel, and a short note in the admin UI and README about what the paid licence adds.

    Open-source companies earn most from teams that already run the software in production. Registration turns anonymous forks into known accounts.

    Risks and how to handle them

    • Community backlash at paywalls

      Keep everything that is open today open, and charge for new enterprise features and support.

    Evidence

    • GitHubmetorial/metorial: 3.4k stars, 371 forks, Apache 2.0, 2,927 commits; the README says "Connect any AI model to 1200+ integrations", that the platform is open source and self-hostable, and that the server index holds "more than 5000 MCP servers". metorial/mcp-index: 281 stars, 41 forks, MIT. The homepage says "1,000+ integrations" and the YC launch post "600+ integrations".Source
    • Your websiteHomepage (checked November 9, 2026): H1 "Put AI to work across your company, safely"; meta description "One MCP URL behind your SSO. 1,000+ integrations, role-based tool access, audit logs and shared skills, in Claude, ChatGPT and Cursor. Open source."; one named customer quote (Moonfire); the FAQ states SOC 2 Type II and GDPR compliance with EU and US regions, and that Metorial can be self-hosted or run on Metorial Cloud.Source
    • Competitor researchRivals' own homepages: Runlayer, "The operating layer for AI", shows eight customer logos, four case studies and quotes from named security leaders, with no public pricing; Composio offers "1,500+ apps" and a free plan with 100,000 tool calls a month; Arcade claims "8,000+ permission-aware tools" and "Free to start. Priced by usage."; Obot is open source, with a free self-hosted Community edition for up to 100 users and a contact-sales Enterprise tier.Source
  • 13

    Metorial publishes 146 scenario pages (such as Salesforce pipeline digests to Teams, or App Store review triage into GitHub), and integration pages already have a "Deploy" link into the platform. Shared skills are a core product feature. A scenario that a reader can't run is a missed activation.

    Medium impactMedium effort First result in 1 month Founding Engineer

    First action

    Package the five most-viewed scenarios as skills.

    How to measure it

    Scenario-to-first-call rate

    Now
    –
    Target
    25%
    By
    2 quarters

    Steps

    1. 1

      Rank scenarios

      By traffic.

      Founder/CEO 1 day

    2. 2

      Package skills

      Top five first.

      Founding Engineer 1 week

    3. 3

      Add the button

      Deep link with integrations pre-selected.

      Founding Engineer 3 days

    4. 4

      Onboarding hook

      Offer scenarios after sign-up.

      Founding Engineer 3 days

    5. 5

      Roll out

      Remaining scenarios in batches.

      Founding Engineer 1 month

    6. 6

      Measure

      Time to first call by scenario.

      Founder/CTO monthly

    First result
    1 month
    Cost
    Engineering time only
    Owner
    Founding Engineer

    Smallest test

    Add one-click run to five scenarios.

    Why it works

    Turn each scenario into a packaged skill with the needed integrations and policies, and add a "Run this scenario" button that opens the platform with everything pre-selected. Track which scenarios get people to their first tool call fastest and feature those on the homepage and in onboarding.

    Templates cut time-to-value, and scenarios already rank for specific workflow searches.

    Risks and how to handle them

    • Scenarios break as integrations change

      Run a nightly test of each packaged skill.

    Evidence

    • Your websiteSitemap (checked November 9, 2026): 128 core pages (including industry pages such as /for/wealth-management and /for/saas, and 23 agent-client pages under /agents/), 38 blog posts, 25 comparison pages, 45 answer pages under /for-ai-crawlers/, 146 scenario pages, 1,000+ integration pages, and 250 integration-by-client setup pages (50 integrations times 5 clients). No pages in languages other than English.Source
    • Your websiteIntegration pages: /integrations/highlevel lists 37 tools with inline setup steps for five clients, but /integrations/highlevel/claude returns a 404, and the title spells the brand "Highlevel" while the changelog writes "HighLevel"; /integrations/salesforce lists 14 tools, links to separate client guides, 15 related scenarios and a "Deploy Salesforce" link into the platform, and carries no customer proof. Titles follow "... MCP Server for Claude & Cursor | Metorial".Source
  • 14

    Metorial shipped ten changelog entries in three weeks (shared skills over Magic MCP, a Usage view, admin setup links, passkey sign-in), but the site has no newsletter signup. Dev users who signed up months ago don't know what changed.

    Medium impactLow effort First result in 2 weeks Founder/CEO

    First action

    Draft the November email from the existing changelog.

    How to measure it

    Dormant accounts with a new tool call within 7 days

    Now
    –
    Target
    5% per send
    By
    90 days

    Steps

    1. 1

      Pick a tool

      Product email from the user database.

      Backend Engineer 1 day

    2. 2

      Draft the template

      Three changes, one integration, one scenario.

      Founder/CEO 1 day

    3. 3

      Segment

      Dev, Scale, Enterprise.

      Backend Engineer 1 day

    4. 4

      Add signup boxes

      Changelog and blog.

      Founding Engineer 2 hours

    5. 5

      Send

      First Tuesday of the month.

      Founder/CEO monthly

    6. 6

      Measure

      Return visits and tool calls.

      Founder/CTO monthly

    First result
    2 weeks
    Cost
    Under $50 a month
    Owner
    Founder/CEO

    Smallest test

    Send one digest to Dev users inactive for 30 days.

    Why it works

    Send a monthly product email to all users: the three biggest changes, one new integration batch, one scenario to try and one customer tip, each linking into the platform. Segment by plan so Dev users see what Scale adds, and add a signup box on the changelog and blog.

    Shipping fast only helps retention if users notice. A monthly digest brings dormant accounts back for almost no effort.

    Risks and how to handle them

    • Unsubscribes

      Keep it short and monthly.

    Evidence

    • Your websiteChangelog: ten entries between October 18 and November 9, 2026, including "Share skills over Magic MCP", "Usage: see what your agents are doing at a glance", "Send admin setup links for Google and Microsoft integrations" and "Create a Metorial organization from the CLI"; 17 updates in 90 days and no newsletter signup found.Source
    • Your websitePricing page: Dev "$0/mo" for "Small projects and personal use" (500,000 tool calls, 2 team members, no production access); Scale "$250/mo" for "Businesses and teams that need more" (2,500,000 tool calls, 20 team members, production access, "basic compliance", support via Discord and email), with SAML SSO and access management marked not included and overage of "$1 per 1,500 additional tool calls"; Enterprise is custom-priced and is the only plan with "SAML, SSO and IAM", "Granular permissions (RBAC)" and "Advanced compliance (SOC 2 II, GDPR)". No free trial, annual option or FAQ; developers are told to "npm install -g @metorial/cli".Source
  • 15

    Every Metorial deal goes through a security review: the product holds employees' credentials for company systems. The site already claims SOC 2 Type II and GDPR, has a trust center, and has an answer page on the "AI agent security questionnaire". Pricing puts "Advanced compliance" only in Enterprise, which leaves buyers unsure what smaller plans get.

    High impactMedium effort First result in 1 month Founder/CTO

    First action

    Fill in one standard questionnaire using the existing security answer page.

    How to measure it

    Sales cycle length

    Now
    –
    Target
    -25%
    By
    2 quarters

    Steps

    1. 1

      Collect documents

      SOC 2, DPA, subprocessors, pen test.

      Founder/CTO 1 week

    2. 2

      Pre-fill the questionnaire

      One common standard.

      Founder/CTO 1 week

    3. 3

      Draw the architecture

      Enclaves, vault, regions, data flow.

      Infrastructure/DevOps Engineer 3 days

    4. 4

      Publish the pack

      On the trust center behind a short form.

      Founding Engineer 2 days

    5. 5

      Clarify by plan

      Pricing row: which documents each plan gets.

      Founder/CEO 1 hour

    6. 6

      Use in deals

      Send with every demo follow-up.

      Founder/CEO ongoing

    First result
    1 month
    Cost
    Founder and engineer time
    Owner
    Founder/CTO

    Smallest test

    Send the pre-filled questionnaire to the next three prospects.

    Why it works

    Build one downloadable pack: SOC 2 report access, a pre-filled standard questionnaire, architecture and data-flow diagrams (enclaves, vault, regions), subprocessors, a pen-test summary and the DPA. Gate it behind a light form that creates a lead, and say plainly on the pricing page which compliance documents each plan gets.

    Security review is often the longest step in the deal. Answering it up front shortens cycles and shows maturity beyond the team's size.

    Risks and how to handle them

    • Answers go out of date

      Review the pack each quarter and date it.

    Evidence

    • Your websiteHomepage (checked November 9, 2026): H1 "Put AI to work across your company, safely"; meta description "One MCP URL behind your SSO. 1,000+ integrations, role-based tool access, audit logs and shared skills, in Claude, ChatGPT and Cursor. Open source."; one named customer quote (Moonfire); the FAQ states SOC 2 Type II and GDPR compliance with EU and US regions, and that Metorial can be self-hosted or run on Metorial Cloud.Source
    • Your websitePricing page: Dev "$0/mo" for "Small projects and personal use" (500,000 tool calls, 2 team members, no production access); Scale "$250/mo" for "Businesses and teams that need more" (2,500,000 tool calls, 20 team members, production access, "basic compliance", support via Discord and email), with SAML SSO and access management marked not included and overage of "$1 per 1,500 additional tool calls"; Enterprise is custom-priced and is the only plan with "SAML, SSO and IAM", "Granular permissions (RBAC)" and "Advanced compliance (SOC 2 II, GDPR)". No free trial, annual option or FAQ; developers are told to "npm install -g @metorial/cli".Source
    • Your websiteAnswer page "How to Connect Salesforce to Claude Securely" (updated August 25, 2026): a five-step CLI setup, a deployment table, five FAQs and cited sources, linking to the Salesforce integration page, Magic MCP, Protoguard and pricing. It sits under /for-ai-crawlers/, one of 45 such answers (for example "ai-agent-security-questionnaire", "sso-and-scim-for-ai-tools" and "roll-out-ai-agents-to-employees").Source
  • 16

    Metorial's pitch starts with identity: SAML SSO, SCIM group sync, and access that "follows users and teams". It just shipped admin setup links for Google and Microsoft integrations. The IT admin who would roll it out often starts in their identity provider's app catalogue.

    Medium impactMedium effort First result in 1-2 quarters Founder/CEO

    First action

    List which identity-provider catalogues already include Metorial.

    How to measure it

    Marketplace-sourced sign-ups

    Now
    –
    Target
    30 a quarter
    By
    2 quarters

    Steps

    1. 1

      Audit listings

      Where Metorial is and is not listed.

      Founder/CEO 1 day

    2. 2

      Build the apps

      SAML and SCIM app templates.

      Backend Engineer 2 weeks

    3. 3

      Write setup pages

      One per identity provider.

      Founding Engineer 1 week

    4. 4

      Submit

      Catalogue reviews.

      Founder/CEO 1 month

    5. 5

      Tag sources

      UTM links per listing.

      Founding Engineer 1 day

    6. 6

      Measure

      Sign-ups by listing.

      Founder/CTO monthly

    First result
    1-2 quarters
    Cost
    Engineering time; some catalogues need a partner account
    Owner
    Founder/CEO

    Smallest test

    Publish one identity-provider setup page and submit one listing.

    Why it works

    Check where Metorial is already listed, then apply to the major identity providers' integration catalogues and the Google Workspace and Microsoft marketplaces with a pre-built SAML and SCIM app. Each listing links to a matching setup page, and the SSO plan (idea 1) makes the listing self-serve.

    Catalogue listings put Metorial where the buyer is already setting up SSO, and they act as a trust signal.

    Risks and how to handle them

    • Long review queues

      Submit to two catalogues at once.

    Builds on

    Evidence

    • Your websiteScale-ups page: Metorial "helps scale-ups roll out AI agents with company SSO and one Magic MCP link", with policies "per user and group"; it names no plan or price and its only call to action is "Book a demo". The SAML SSO feature page also names no plan or price ("Book a demo, or talk to sales").Source
    • Your websiteChangelog: ten entries between October 18 and November 9, 2026, including "Share skills over Magic MCP", "Usage: see what your agents are doing at a glance", "Send admin setup links for Google and Microsoft integrations" and "Create a Metorial organization from the CLI"; 17 updates in 90 days and no newsletter signup found.Source
  • 17

    The pricing page and the answer guides tell people to install the CLI and run "metorial integrations setup", yet @metorial/cli gets about 11 weekly downloads while the SDK and utility packages total about 9,600. Either people install it some other way, or the guides' main call to action is not working.

    Medium impactLow effort First result in 3 weeks Founding Engineer

    First action

    Compare CLI logins with npm downloads to see how users really install it.

    How to measure it

    New CLI logins per week

    Now
    –
    Target
    3x
    By
    60 days

    Steps

    1. 1

      Measure installs

      Downloads by channel compared with logins.

      Founder/CTO 1 day

    2. 2

      One-line install

      Copy button on every guide.

      Founding Engineer 2 days

    3. 3

      No-install path

      npx or hosted option.

      Founding Engineer 1 week

    4. 4

      End with the URL

      Magic MCP URL printed and copied.

      Founding Engineer 2 days

    5. 5

      Track activation

      Login to first tool call.

      Backend Engineer 2 days

    6. 6

      Review

      Drop-off by step.

      Founder/CTO monthly

    First result
    3 weeks
    Cost
    Engineering time only
    Owner
    Founding Engineer

    Smallest test

    Add an npx one-liner to the Salesforce guide and watch logins.

    Why it works

    Find out how people actually install it (npm, other package managers, binaries). Then make install a single copy-paste line on every integration and guide page, add a no-install path (npx or a hosted shell), and end each CLI flow with the Magic MCP URL ready to paste. Track CLI logins as an activation event.

    If the CLI is the shortest path to a first governed tool call, removing friction there raises activation across every guide at once.

    Risks and how to handle them

    • Download counts miss other channels

      Use logins as the real measure.

    Evidence

    • ResearchPackage registries: 29 npm packages with about 9,600 weekly downloads, mostly SDK and utility packages, while @metorial/cli gets 11 a week; the Python package "metorial" is at version 2.7.0, released November 9, 2026.Source
    • Your websiteAnswer page "How to Connect Salesforce to Claude Securely" (updated August 25, 2026): a five-step CLI setup, a deployment table, five FAQs and cited sources, linking to the Salesforce integration page, Magic MCP, Protoguard and pricing. It sits under /for-ai-crawlers/, one of 45 such answers (for example "ai-agent-security-questionnaire", "sso-and-scim-for-ai-tools" and "roll-out-ai-agents-to-employees").Source
    • Your websitePricing page: Dev "$0/mo" for "Small projects and personal use" (500,000 tool calls, 2 team members, no production access); Scale "$250/mo" for "Businesses and teams that need more" (2,500,000 tool calls, 20 team members, production access, "basic compliance", support via Discord and email), with SAML SSO and access management marked not included and overage of "$1 per 1,500 additional tool calls"; Enterprise is custom-priced and is the only plan with "SAML, SSO and IAM", "Granular permissions (RBAC)" and "Advanced compliance (SOC 2 II, GDPR)". No free trial, annual option or FAQ; developers are told to "npm install -g @metorial/cli".Source

Big bets 3

Larger plays that can change the trajectory.

  • 18

    Metorial's homepage names the problem as "shadow MCP servers", "API keys in config files" and "no record of what ran". A security lead can't buy a fix for a problem they can't see. Today nothing lets them measure it in their own company.

    High impactHigh effort First result in 1 quarter Founder/CTO

    First action

    Prototype the scan for the three most common AI clients' config files.

    How to measure it

    Scanner-to-sign-up rate

    Now
    –
    Target
    10%
    By
    2 quarters

    Steps

    1. 1

      Define the checks

      Config locations, static keys, server count.

      Founder/CTO 1 week

    2. 2

      Build locally-only

      No secrets leave the machine.

      Founding Engineer 3 weeks

    3. 3

      Team summary

      Opt-in aggregate counts.

      Backend Engineer 2 weeks

    4. 4

      Migration command

      Import found servers into Metorial.

      Founding Engineer 2 weeks

    5. 5

      Security review

      Independent review of the scanner itself.

      Infrastructure/DevOps Engineer 1 week

    6. 6

      Launch

      GitHub, Hacker News and the blog.

      Founder/CEO 1 week

    7. 7

      Measure

      Runs, team summaries, sign-ups.

      Founder/CTO monthly

    First result
    1 quarter
    Cost
    About two engineer-months
    Owner
    Founder/CTO

    Smallest test

    Publish a one-page script that counts configured servers and keys, and see who shares it.

    Why it works

    Ship an open-source scanner (CLI plus an MDM-friendly script) that reads local AI-client MCP configs on a machine and reports, locally and without sending secrets, how many servers are configured, which hold static keys, and which tools they expose. A team summary view (opt-in, counts only) ends with "Move these behind one governed URL" and a migration command.

    Free diagnostic tools create urgency and qualified leads. The output is a migration plan for Metorial, and open source fits the brand.

    Risks and how to handle them

    • A security tool must itself be trusted

      Open source, local-only by default, reviewed before launch.

    Evidence

    • Your websiteHomepage (checked November 9, 2026): H1 "Put AI to work across your company, safely"; meta description "One MCP URL behind your SSO. 1,000+ integrations, role-based tool access, audit logs and shared skills, in Claude, ChatGPT and Cursor. Open source."; one named customer quote (Moonfire); the FAQ states SOC 2 Type II and GDPR compliance with EU and US regions, and that Metorial can be self-hosted or run on Metorial Cloud.Source
    • GitHubmetorial/metorial: 3.4k stars, 371 forks, Apache 2.0, 2,927 commits; the README says "Connect any AI model to 1200+ integrations", that the platform is open source and self-hostable, and that the server index holds "more than 5000 MCP servers". metorial/mcp-index: 281 stars, 41 forks, MIT. The homepage says "1,000+ integrations" and the YC launch post "600+ integrations".Source
  • 19

    Metorial maintains an MCP index of "more than 5000 MCP servers" (281 stars on GitHub), but it lives as a repo folder, while the website covers only the 1,000+ managed integrations. People searching for an MCP server for a niche tool find GitHub lists, not Metorial.

    High impactHigh effort First result in 1 quarter Founding Engineer

    First action

    Generate pages for the 200 most-starred servers in the index.

    How to measure it

    Directory-sourced sign-ups

    Now
    –
    Target
    200 a month
    By
    2 quarters

    Steps

    1. 1

      Normalise the data

      Name, repo, tools, licence, activity.

      Backend Engineer 2 weeks

    2. 2

      Build page templates

      Searchable directory and server pages.

      Founding Engineer 2 weeks

    3. 3

      Add run buttons

      Into custom or Docker MCP in the platform.

      Founding Engineer 1 week

    4. 4

      Claim and badge

      Maintainer verification and README badge.

      Founding Engineer 2 weeks

    5. 5

      Launch

      Announce on GitHub and the blog.

      Founder/CEO 1 week

    6. 6

      Measure

      Traffic, runs and sign-ups.

      Founder/CTO monthly

    First result
    1 quarter
    Cost
    About two engineer-months
    Owner
    Founding Engineer

    Smallest test

    Publish 50 directory pages and measure impressions.

    Why it works

    Publish the index as metorial.com pages: one per server with what it does, its tools, maintenance signals (last commit, licence), a basic safety summary, and a button to run it as a custom or Docker MCP server behind Metorial's gateway. Let maintainers claim their page, and add a badge for their README.

    A directory captures long-tail search, and maintainer badges create backlinks. Every entry is a reason to add governance to a server the reader already wants.

    Risks and how to handle them

    • Listing an unsafe server looks like an endorsement

      Label community entries clearly, and run them only behind policies and sandboxes.

    Evidence

    • GitHubmetorial/metorial: 3.4k stars, 371 forks, Apache 2.0, 2,927 commits; the README says "Connect any AI model to 1200+ integrations", that the platform is open source and self-hostable, and that the server index holds "more than 5000 MCP servers". metorial/mcp-index: 281 stars, 41 forks, MIT. The homepage says "1,000+ integrations" and the YC launch post "600+ integrations".Source
    • Your websiteSitemap (checked November 9, 2026): 128 core pages (including industry pages such as /for/wealth-management and /for/saas, and 23 agent-client pages under /agents/), 38 blog posts, 25 comparison pages, 45 answer pages under /for-ai-crawlers/, 146 scenario pages, 1,000+ integration pages, and 250 integration-by-client setup pages (50 integrations times 5 clients). No pages in languages other than English.Source
  • 20

    Metorial sees what few others can: which integrations companies connect, which tools agents call, how often calls are blocked by policy, and cost per session (the new Usage view). The press calls its evidence "largely company-reported", and the blog already writes about cost per completed task and pilots that stall.

    High impactHigh effort First result in 2 quarters Founder/CEO

    First action

    Check which aggregated metrics the platform can export with no customer identifiers.

    How to measure it

    Report-sourced sign-ups

    Now
    –
    Target
    300
    By
    first report + 60 days

    Steps

    1. 1

      Privacy rules

      Aggregation thresholds and opt-out.

      Founder/CTO 1 week

    2. 2

      Pick metrics

      Eight to ten clear charts.

      Founder/CEO 1 week

    3. 3

      Export data

      Aggregate pipeline.

      Backend Engineer 2 weeks

    4. 4

      Write the report

      Findings and method.

      Founder/CEO 3 weeks

    5. 5

      Pitch press

      Exclusive to one trade publication.

      Founder/CEO 2 weeks

    6. 6

      Benchmark calls

      Customers see where they stand.

      Founder/CEO ongoing

    First result
    2 quarters
    Cost
    Founder time plus design
    Owner
    Founder/CEO

    Smallest test

    Publish one chart (most-connected tools) as a blog post.

    Why it works

    Publish an annual report from aggregated, anonymised platform data (opt-out for customers): the most-connected tools by team, read-versus-write ratios, policy blocks, prompt-injection detections and cost per completed task. Release the charts openly, brief trade press, and invite customers to a benchmark call comparing their numbers with the median.

    Original data earns coverage and links that a small team can't buy, and it turns "company-reported" claims into an industry reference.

    Risks and how to handle them

    • Customer privacy concerns

      Only aggregates over strict minimum counts, with opt-out and a published method.

    Evidence

    • Your websiteChangelog: ten entries between October 18 and November 9, 2026, including "Share skills over Magic MCP", "Usage: see what your agents are doing at a glance", "Send admin setup links for Google and Microsoft integrations" and "Create a Metorial organization from the CLI"; 17 updates in 90 days and no newsletter signup found.Source
    • PressRuntimeWire startup spotlight (October 2026): Metorial's "public evidence of demand remains early and largely company-reported"; the company is "betting that as companies give agents access to real systems, identity, permissions and audit trails become shared infrastructure."Source
    • Your websiteBlog: 38 posts, recently about one a week, with slugs such as "ai-agent-policy-on-one-page" (November 5, 2026), "whose-permissions-does-an-agent-use", "cost-per-completed-task" and "who-owns-an-agent-after-pilot".Source

Wildcards 3

Unexpected experiments, each with a cheap way to test it.

  • 21

    Metorial is based in Vienna, offers an EU region and GDPR compliance, and sells governance, which European buyers weigh heavily. Yet every page is in English, and the US-focused rivals do the same.

    Medium impactMedium effort First result in 1 quarter Founder/CEO

    First action

    Count current sign-ups and demos from German-speaking countries.

    How to measure it

    DACH demo requests

    Now
    –
    Target
    +50%
    By
    2 quarters

    Steps

    1. 1

      Size the base

      DACH sign-ups and demos so far.

      Founder/CEO 1 day

    2. 2

      Translate core pages

      Native review, not raw machine output.

      Founder/CEO 2 weeks

    3. 3

      Add hreflang

      en and de pairs.

      Founding Engineer 1 day

    4. 4

      Lead with EU

      EU region and GDPR first.

      Founder/CEO 2 days

    5. 5

      Local outreach

      Meetups and IT leader events.

      Founder/CEO 1 quarter

    6. 6

      Compare

      Conversion against English pages.

      Founder/CTO monthly

    First result
    1 quarter
    Cost
    Translation review time
    Owner
    Founder/CEO

    Smallest test

    Publish a German security page and send it to ten DACH leads.

    Why it works

    Translate the homepage, pricing, security, scale-ups and enterprise pages, plus the five strongest answer guides, into German with hreflang. Lead with EU hosting, GDPR and self-hosting. Run a small test of founder-led outreach and events in Vienna, Munich and Zurich.

    Native-language governance content is scarce in this category, and a local team is credible to European IT and works councils.

    Risks and how to handle them

    • Keeping translations current

      Translate only the slow-changing core pages.

    Evidence

    • HiringY Combinator jobs (open now): Founding Engineer ($100K-$180K, 1-2% equity), Backend Engineer and Infrastructure/DevOps Engineer, all remote; the company is YC F25, founded 2025, a team of 4, based in Vienna. The posts ask for "deep distributed systems knowledge", TypeScript and Go, "Good understanding of AI agents (ideally MCP and tool calls)" and "AWS + HashiStack". The careers page sits on meteor.computer under the name "Meteor Computing".Source
    • Your websiteSitemap (checked November 9, 2026): 128 core pages (including industry pages such as /for/wealth-management and /for/saas, and 23 agent-client pages under /agents/), 38 blog posts, 25 comparison pages, 45 answer pages under /for-ai-crawlers/, 146 scenario pages, 1,000+ integration pages, and 250 integration-by-client setup pages (50 integrations times 5 clients). No pages in languages other than English.Source
    • Your websiteHomepage (checked November 9, 2026): H1 "Put AI to work across your company, safely"; meta description "One MCP URL behind your SSO. 1,000+ integrations, role-based tool access, audit logs and shared skills, in Claude, ChatGPT and Cursor. Open source."; one named customer quote (Moonfire); the FAQ states SOC 2 Type II and GDPR compliance with EU and US regions, and that Metorial can be self-hosted or run on Metorial Cloud.Source
  • 22

    Real visitors already pass Core Web Vitals (75th-percentile LCP 1.8 s, INP 66 ms), so speed is not urgent. But Google's phone test shows a 9.2 s lab LCP and about 256 KB of unused JavaScript, and lists four accessibility issues (unnamed buttons, contrast, heading order, touch targets). The same template will serve thousands of integration and directory pages, often to first-time visitors on slower devices.

    Low impactLow effort First result in 2 weeks Founding Engineer

    First action

    Fix button labels and contrast in the shared components.

    How to measure it

    Integration-page mobile LCP (lab)

    Now
    –
    Target
    under 4 s
    By
    60 days

    Steps

    1. 1

      Profile the template

      Which bundles integration pages load.

      Founding Engineer 1 day

    2. 2

      Split code

      Load interactive mockups only where used.

      Founding Engineer 3 days

    3. 3

      Cache assets

      Long lifetimes on static files.

      Infrastructure/DevOps Engineer 1 day

    4. 4

      Fix accessibility

      Labels, contrast, headings, touch targets.

      Founding Engineer 2 days

    5. 5

      Retest

      PageSpeed on five templates.

      Founding Engineer 1 hour

    6. 6

      Guard

      Bundle-size check in CI.

      Infrastructure/DevOps Engineer 1 day

    First result
    2 weeks
    Cost
    Engineering time only
    Owner
    Founding Engineer

    Smallest test

    Fix the accessibility issues this week.

    Why it works

    Split the JavaScript that the shared marketing template loads so integration and guide pages ship only what they use, set long cache lifetimes on static assets, and fix the four accessibility issues in the shared components. Enterprise buyers often check accessibility during procurement.

    One template fix improves every current and future programmatic page.

    Risks and how to handle them

    • Effort with little visible gain

      Timebox it to one week, since field data already passes.

    Evidence

    • Google PageSpeedHomepage, November 10, 2026: mobile score 64 (desktop 85) and a lab LCP of 9.2 s on a phone, but real visitors pass Core Web Vitals (75th-percentile LCP 1.8 s, 30% faster than before, INP 66 ms, CLS 0). About 256 KB of unused JavaScript and four accessibility issues: buttons without accessible names, low contrast, heading order and small touch targets.Source
    • Your websiteSitemap (checked November 9, 2026): 128 core pages (including industry pages such as /for/wealth-management and /for/saas, and 23 agent-client pages under /agents/), 38 blog posts, 25 comparison pages, 45 answer pages under /for-ai-crawlers/, 146 scenario pages, 1,000+ integration pages, and 250 integration-by-client setup pages (50 integrations times 5 clients). No pages in languages other than English.Source
  • 23

    Scale support runs on Discord and email, and the open-source repo has 3.4k stars, but the site links no public community. Dev users and self-hosters have nowhere to share skills, ask questions or meet the team.

    Medium impactLow effort First result in 1 month Founder/CTO

    First action

    Create public channels and link them from the README and docs.

    How to measure it

    Active members

    Now
    –
    Target
    300
    By
    2 quarters

    Steps

    1. 1

      Set up channels

      Skills, integrations, self-host, admins.

      Founder/CTO 1 day

    2. 2

      Link everywhere

      Docs, README, changelog, app.

      Founding Engineer 1 day

    3. 3

      Office hours

      Weekly, 30 minutes.

      Founder/CTO weekly

    4. 4

      Skill showcase

      Members share skills.

      Founder/CEO weekly

    5. 5

      Answers to docs

      Promote good threads into guides.

      Founding Engineer monthly

    6. 6

      Measure

      Active members and retention of members compared with others.

      Founder/CTO monthly

    First result
    1 month
    Cost
    $0
    Owner
    Founder/CTO

    Smallest test

    Open one public #skills channel and invite recent sign-ups.

    Why it works

    Open a public community with channels for skills, integrations, self-hosting and admins rolling out to a company. Founders hold weekly office hours, and good answers move into the docs or answer guides. Link it from docs, the README, the changelog and the product.

    Communities keep free and open-source users engaged and surface feature requests and champions. They also take repeat questions off a four-person team.

    Risks and how to handle them

    • Moderation load

      Set clear rules and limit channels.

    Evidence

    • Your websitePricing page: Dev "$0/mo" for "Small projects and personal use" (500,000 tool calls, 2 team members, no production access); Scale "$250/mo" for "Businesses and teams that need more" (2,500,000 tool calls, 20 team members, production access, "basic compliance", support via Discord and email), with SAML SSO and access management marked not included and overage of "$1 per 1,500 additional tool calls"; Enterprise is custom-priced and is the only plan with "SAML, SSO and IAM", "Granular permissions (RBAC)" and "Advanced compliance (SOC 2 II, GDPR)". No free trial, annual option or FAQ; developers are told to "npm install -g @metorial/cli".Source
    • GitHubmetorial/metorial: 3.4k stars, 371 forks, Apache 2.0, 2,927 commits; the README says "Connect any AI model to 1200+ integrations", that the platform is open source and self-hostable, and that the server index holds "more than 5000 MCP servers". metorial/mcp-index: 281 stars, 41 forks, MIT. The homepage says "1,000+ integrations" and the YC launch post "600+ integrations".Source
More example playbooks and reviews
ProgrammableDigital

Show Metorial’s product working

Turn screenshots or the website into an animated product demo in its own brand, for the homepage, launch posts and outreach in this playbook.

Make a demo (opens programmable.digital in a new tab)

More growth playbooks

Other companies we have researched, each read from what they are hiring for.

  • CareSwift growth strategyAI-assisted EMS narrative software for agencies. Crews enter call details, answer clarifying questions built from the agency's own…23 growth ideas
  • Maesa growth strategyA beauty company that creates, incubates and runs brands in haircare, fragrance, body and wellness (Kristin Ess, Hairitage, Fine'ry, Being…21 growth ideas
  • Measure Marketer growth strategyA Hyderabad tracking and measurement agency that sets up and fixes GA4, GTM, Meta CAPI, Google Ads conversions, server-side tagging…21 growth ideas
  • Armature growth strategyAgent experience for developer tools: a done-with-you discoverability service that measures how coding agents choose tools in a category…22 growth ideas
  • Rubrik growth strategyA security and AI operations company: Rubrik Security Cloud protects, monitors and recovers data, identities and workloads across clouds…22 growth ideas
  • France Luxe growth strategyA women-owned hair accessories company selling French-made cellulose acetate clips, barrettes, combs and headbands under France Luxe, and…20 growth ideas
  • Loyal growth strategyA clinical-stage veterinary medicine company developing drugs intended to extend the healthy lifespan of dogs (LOY-002 for senior dogs…21 growth ideas
  • NetDocuments growth strategyA cloud document, email and knowledge management platform for legal work, with the legal context graph, ndMAX legal AI apps, MCP…20 growth ideas

See every playbook and website review

Independent research. Programmable wrote this playbook from public sources, as they stood on November 10, 2026, and cites them throughout. Metorial did not commission, review or endorse it, and Programmable is not affiliated with Metorial. Company and product names are trademarks of their owners, used only to identify them. The ideas are suggestions and opinion, not professional, legal or financial advice, and not statements about how Metorial runs its business.

Something wrong, or do you represent Metorial? Email [email protected] and we will correct it or take it down.

30/60/90 roadmap

What to focus on in each month, and the ideas that carry it.

  1. Days 1 to 307 ideas

    Make SSO buyable, then quick fixes for proof and consistency: the demo page, the integration count and names, the comparison check, the cost calculator, answer-page links and industry stories.

  2. Days 31 to 6010 ideas

    Catalogue setup pages, quantified stories, the Dev-to-production moment, product-qualified leads, the self-host upgrade path, one-click scenarios, the changelog email, the security pack, identity marketplaces and CLI adoption.

  3. Days 61 to 906 ideas

    Commit to the big bets (shadow-MCP scanner, public server directory, annual usage report) and test the wildcards: German pages, a lighter integration template and a public community.

The research behind the plan

3 research areas, each summarised in a few lines. Open any of them for the full detail.

Customers

What customers say

Customers value one place for all data sources and tools. The most cited problem: shadow MCP servers and API keys in config files.

  • Shipping customer connector requests without building OAuth and MCP plumbing
  • No record of what agents did, or for whom
  • In their words: “shadow MCP”

Market

Competitive landscape

5 competitors compared: Runlayer, Composio, Arcade, Obot and 1 more.

  • Gap to exploit: SSO-governed AI access at a published price for mid-sized companies
  • Runlayer: No public pricing or self-serve path
  • Composio: No named customers on the homepage

Hiring signals

Metorial is a four-person YC F25 team and is hiring only engineers: a Founding Engineer, a Backend Engineer and an Infrastructure/DevOps Engineer, all remote, asking for distributed systems, TypeScript and Go, MCP and…

  • Founding Engineer: Distributed systems, TypeScript and Go, MCP; owns product-led growth builds. Ideas 2, 10, 17, 18, 19.
  • Backend Engineer: Microservices and serverless; usage signals and data exports. Ideas 11, 20.
  • Infrastructure/DevOps Engineer: "Experience with large production systems"; self-hosted and on-prem deployments. Ideas 12, 15.

From the team that wrote this playbook

All products
  • ProgrammableAgencyWorking AI systems in days, not quarters (opens programmable.agency in a new tab)
  • ProgrammableInfoAn agent that finds, researches and asks first (opens programmable.info in a new tab)
  • ProgrammableDigitalScreenshots in, an animated product demo out (opens programmable.digital in a new tab)
  • ProgrammableWatchKnow the day a supplier falls out of compliance (opens programmable.watch in a new tab)
  • ProgrammableTalkGet registrants for your events, and know who’s coming (opens programmable.talk in a new tab)
  • ProgrammablePageText what you need, your customer gets a page that works (opens programmable.page in a new tab)
  • ProgrammableCommunityForums with labelled AI members, never empty (opens programmable.community in a new tab)
ProgrammableMarketing
ExamplesLeaderboardsProductsTermsPrivacyCookiesDPALegalContact

© 2026 AI-Powered Growth Intelligence