ProgrammableMarketing/openrouter.ai (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 · 11
  • Big bets · 3
  • Wildcards · 2
  • How we can help
Generated Oct 23, 2026 1 view
  1. Programmable
  2. Playbooks
  3. OpenRouter
OpenRouter logo

growth playbook · Oct 23, 2026

openrouter.ai (opens in a new tab)

OpenRouter growth playbook

Diagnosis

OpenRouter is the leading neutral AI gateway, with huge self-serve reach (more than 10 million users, millions of weekly SDK downloads) and a new parent in Stripe.

Its hiring shows the next step: an enterprise motion with product marketing, BDRs, strategic AEs and GRC; a growth director with a $5M+ paid budget and AEO/SEO; a community lead; Connect, a platform for platforms; and an APAC footprint.

The gaps sit where enterprise buyers decide: the enterprise page has logos but no customer results, the best customer story is split across three URLs, rivals publish comparisons with OpenRouter while OpenRouter has none, and public numbers differ by page. This playbook fixes proof and pricing clarity first, then builds product-signal outbound, an advisory board, Connect and provider onboarding, and longer bets on Stripe distribution and recurring data reports.

Growth Score

8/10

Biggest Opportunity

Turn self-serve scale into enterprise and partner revenue: proof on the enterprise page, product signals for sales, and Connect partners drawn from apps already routing through OpenRouter.

Biggest Risk

Portkey, LiteLLM, Vercel, Cloudflare and Requesty all sell gateways, and some cloud and model providers offer their own; without clear proof and comparisons, enterprise buyers may default to a gateway inside their existing vendor.

01

Start here

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

  1. 1

    Put named customer results on the enterprise page, starting with Descript's "1 week to 2 hours"

    Add the Descript result and link to the enterprise page this week.

    High impact First result in 2 weeks
  2. 2

    Merge the three Descript stories into one page that ranks and gets cited

    Set the canonical on the blog case-study copy.

    Medium impact First result in 1 month
  3. 3

    Publish "OpenRouter vs" gateway pages and migration guides, because rivals already write them

    Pick the three gateways that come up most in deals.

    High impact First result in 1 month

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.

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

    1 of these are things Programmable can build and run for you.

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 · 2Big bets · 10Fill-ins · 6Reconsider · 5
Effort →

Showing 23 of 23 ideas

This week 7

Quick wins with a clear first action.

  • 01

    The new Enterprise PMM is hired to build "customer proof that help our biggest deals close". Yet the enterprise page, where every demo request starts, has eight customer logos and not one customer story, quote or named result, and it does not link to the Descript story, which already proves the enterprise case: 13 models in production through one integration, and new-model evaluation cut from over a week to one or two hours. The only number on the page is an illustrative example.

    High impactLow effort First result in 2 weeks Product Marketing Manager

    First action

    Add the Descript result and link to the enterprise page this week.

    How to measure it

    Enterprise page demo-request rate

    Now
    –
    Target
    +20%
    By
    60 days

    Steps

    1. 1

      Pull the proof

      Descript and Pax Historia numbers plus any approved quotes.

      Product Marketing Manager, Enterprise 1 day

    2. 2

      Ask logo customers

      Request one quote and one number from each of the eight logos.

      Product Marketing Manager, Enterprise 1 week

    3. 3

      Design the band

      Proof under the hero, one line per buyer group.

      Product Designer 2 days

    4. 4

      Ship it

      Enterprise page, demo confirmation email and AE deck.

      Product Marketing Manager, Enterprise 2 days

    5. 5

      Replace the example

      A real blended-cost result in place of the illustrative -27%.

      Product Marketing Manager, Enterprise 1 month

    6. 6

      Measure

      Demo requests per visit before and after.

      Director, Growth Marketing 4 weeks

    First result
    2 weeks
    Cost
    Staff time only
    Owner
    Product Marketing Manager, Enterprise

    Smallest test

    Add the Descript numbers and link under the hero for two weeks and compare demo requests.

    Why it works

    Add a proof band under the hero: the Descript numbers with a link to the story, two short quotes from logo customers, and one line per buyer group (engineering: day-0 models; finance: spend by team; security: zero data retention; procurement: one contract). Swap the illustrative -27% for a real, approved customer figure as soon as one is cleared. Put the same proof in the demo confirmation email and the AE deck.

    Enterprise buyers want to see a company like theirs succeed, and the strongest proof OpenRouter has is published but sits two clicks away from the page that asks for the demo.

    Ready to use

    ChecklistEnterprise proof band

    [ ] Descript: 1 -> 13 models in production, 1+ week -> 1-2 hours to evaluate a new model (link to story)
    [ ] Two quotes from logo customers, with name and title
    [ ] One proof line each for engineering, finance, security and procurement
    [ ] A real blended-cost result to replace the illustrative example
    [ ] Same band in the demo confirmation email
    [ ] Same slide in the AE and SE deck
    [ ] Demo-request rate tracked before and after

    Risks and how to handle them

    • Customers slow to approve quotes

      Start with the published Descript story and add quotes as they clear.

    • Logos without stories look thin

      Rotate in each new story as the advocacy program produces it.

    Evidence

    • HiringProduct Marketing Manager, Enterprise (posted September 29, 2026): owns the enterprise journey "from demo request to signed contract to expansion"; will "build the sales narrative and customer proof that help our biggest deals close", "Build and maintain competitor comparisons", "run win/loss analysis", and "Stand up and scale advocacy, case studies, references, and a customer advisory board or design partner program".Source
    • Your websiteEnterprise page (checked October 22, 2026): H1 "Your inference control plane"; a strip of eight customer logos (Lovable, Replit, Amplitude, Nvidia, Zoom, WTW, NIST, AMD) but no customer story, quote or named result anywhere on the page and no link to a case study; the only cost figure is an illustrative "Example: blended cost per 1M tokens -27%"; calls to action "Book a demo" and "Talk to us".Source
    • Your websiteCustomer proof on the site: two customer stories, /customers/descript ("1 -> 13 Models in production", "1+ week -> 1-2 hours To evaluate a promising new model") and /customers/paxhistoria ("35k" daily players, "4.8M+" monthly requests); the blog's case-study section adds Rasp, an agent used by OpenRouter's own sales team. The Descript story is published at three URLs, each with its own canonical: /customers/descript, /blog/case-studies/descript-case-study/ and /blog/insights/descript-model-evaluation-queue/.Source
  • 02

    The Descript story lives at /customers/descript, /blog/case-studies/descript-case-study/ and /blog/insights/descript-model-evaluation-queue/, and each page names itself as the original. Search engines and AI assistants split the credit three ways, and buyers who find one copy miss the others.

    Medium impactLow effort First result in 1 month Director

    First action

    Set the canonical on the blog case-study copy.

    How to measure it

    Impressions for the main story URL

    Now
    –
    Target
    +50%
    By
    60 days

    Steps

    1. 1

      Pick the main URL

      /customers/descript.

      Director, Growth Marketing 1 hour

    2. 2

      Canonical or redirect

      Blog case-study copy points to it.

      Web Engineer 2 hours

    3. 3

      Rewrite the insights post

      A how-to on evaluating new models, linking the story.

      Content team 2 days

    4. 4

      Update internal links

      Enterprise page, blog index and docs.

      Director, Growth Marketing 1 day

    5. 5

      Set the rule

      One URL per customer story.

      Director, Growth Marketing 1 hour

    6. 6

      Check

      Search Console impressions by URL.

      Director, Growth Marketing 4 weeks

    First result
    1 month
    Cost
    $0
    Owner
    Director, Growth Marketing

    Smallest test

    Add the canonical today and watch Search Console.

    Why it works

    Keep /customers/descript as the main page, point the case-study copy to it with a canonical or redirect, and turn the insights post into a distinct how-to on fast model evaluation that links to the story. Use the same rule for every future customer story.

    One strong URL collects the links and citations that three copies split, and it is the page the enterprise team will link from now on.

    Risks and how to handle them

    • Short-term ranking dip

      Use canonicals first and redirect after the main page is indexed.

    Evidence

    • Your websiteCustomer proof on the site: two customer stories, /customers/descript ("1 -> 13 Models in production", "1+ week -> 1-2 hours To evaluate a promising new model") and /customers/paxhistoria ("35k" daily players, "4.8M+" monthly requests); the blog's case-study section adds Rasp, an agent used by OpenRouter's own sales team. The Descript story is published at three URLs, each with its own canonical: /customers/descript, /blog/case-studies/descript-case-study/ and /blog/insights/descript-model-evaluation-queue/.Source
    • HiringDirector, Growth Marketing: "More than 10 million users and 250k+ apps process 200T+ tokens on the platform every month, nearly all of it starting in self-serve." The role owns "signing up, buying their first credits, and growing their usage", paid media "with a paid budget of $5M+ for 2027", "the AEO/SEO roadmap for openrouter.ai", conversion optimization on "homepage, model pages, pricing, signup, credit purchase", and "incrementality and holdout testing".Source
  • 03

    Requesty publishes a page comparing itself with OpenRouter and blog posts on OpenRouter alternatives, and Portkey has an "OpenRouter alternatives" page. OpenRouter has 3,055 model-vs-model pages but none comparing itself with Portkey, LiteLLM, Vercel AI Gateway, Requesty or Cloudflare AI Gateway, and no guide for moving from another gateway. Buyers searching those terms read the rivals' version.

    High impactMedium effort First result in 1 month Product Marketing Manager

    First action

    Pick the three gateways that come up most in deals.

    How to measure it

    Comparison-page conversions

    Now
    –
    Target
    150 a month
    By
    90 days

    Steps

    1. 1

      Pick rivals

      Three most common in win/loss notes.

      Product Marketing Manager, Enterprise 1 day

    2. 2

      Check facts

      Each rival's current features and pricing, dated.

      Product Marketing Manager, Enterprise 3 days

    3. 3

      Write pages

      Fair tables plus switching steps.

      Product Marketing Manager, Enterprise 1 week

    4. 4

      Write migration guides

      Code diffs per rival.

      Solutions Architect 1 week

    5. 5

      Legal review

      Claims and trademarks.

      Legal 3 days

    6. 6

      Refresh

      Quarterly, with dates on each page.

      Product Marketing Manager, Enterprise quarterly

    First result
    1 month
    Cost
    Staff time only
    Owner
    Product Marketing Manager, Enterprise

    Smallest test

    Publish one comparison and one migration guide.

    Why it works

    Turn the PMM's competitor comparisons into fair public pages: models and providers, routing and fallbacks, pricing and fees, data policies and regions, enterprise controls, and switching steps. Add a migration guide per rival with code diffs, and link them from the enterprise page.

    Comparison searches come from buyers who are already choosing, and a fair page from OpenRouter answers them better than a rival's.

    Risks and how to handle them

    • Out-of-date claims

      Date every page and cite each rival's own docs.

    Evidence

    • HiringProduct Marketing Manager, Enterprise (posted September 29, 2026): owns the enterprise journey "from demo request to signed contract to expansion"; will "build the sales narrative and customer proof that help our biggest deals close", "Build and maintain competitor comparisons", "run win/loss analysis", and "Stand up and scale advocacy, case studies, references, and a customer advisory board or design partner program".Source
    • Your websiteSitemaps (read October 22, 2026): 1,196 model pages, 3,055 model-vs-model compare pages, 527 docs pages (including model migration guides such as Claude 5.5 and GPT-6), 159 blog URLs, 16 collections, 9 rankings and benchmark pages, a /providers/apply form and 53 "Works with OpenRouter" pages; no page comparing OpenRouter with other AI gateways, no guide for moving from another gateway, and no OpenRouter Connect page.Source
    • Competitor researchRivals write about OpenRouter on their own sites: Requesty publishes /vs/openrouter and posts such as "openrouter-eu-alternative-european-ai-gateway" and "best-llm-routing-platforms-compared-2026-requesty-portkey-litellm-openrouter"; Portkey publishes /alternatives/openrouter-alternatives.Source
    • Competitor researchRivals' own homepages: Portkey, "AI gateway, observability, guardrails, and prompt management"; LiteLLM, "Open-Source AI Gateway & LLM Proxy", "The AI Gateway for platform teams"; Vercel AI Gateway, "400+ AI models"; Requesty, "600+ models" with "EU data residency and enterprise governance"; Cloudflare AI Gateway, an "AI Application Control Plane" with "unified billing".Source
  • 04

    The homepage says "700T+ Monthly Tokens", "500+ Models" and "10M+ Global Users"; the Stripe announcement says "10+ trillion tokens per day from 400+ AI models"; the growth-marketing post says "200T+ tokens ... every month"; and the homepage also says apps reach "4.2M+ users". Each is true at some date, but together they read as inconsistent to a buyer or a journalist.

    Medium impactLow effort First result in 2 weeks Director

    First action

    List every public figure and where it appears.

    How to measure it

    Pages with the current numbers

    Now
    –
    Target
    100% of key pages
    By
    30 days

    Steps

    1. 1

      Inventory

      Every figure on site, blog, brand page and jobs.

      Director, Growth Marketing 1 day

    2. 2

      Define each

      Tokens per day or month, users vs apps.

      Data team 1 day

    3. 3

      Publish the page

      With dates and definitions.

      Director, Growth Marketing 1 day

    4. 4

      Update pages

      Homepage, enterprise, brand kit, job posts.

      Web Engineer 1 day

    5. 5

      Set a cadence

      Monthly refresh.

      Director, Growth Marketing monthly

    6. 6

      Brief the team

      Sales and recruiting use the page.

      Product Marketing Manager, Enterprise 1 hour

    First result
    2 weeks
    Cost
    $0
    Owner
    Director, Growth Marketing

    Smallest test

    Fix the job-post figure today.

    Why it works

    Publish a short "OpenRouter by the numbers" page with each figure, its definition and the date, update it monthly, and point the homepage, enterprise page, press kit and job posts at it.

    Consistent, dated numbers are what journalists and AI assistants quote, and they stop buyers wondering which figure is right.

    Risks and how to handle them

    • Numbers move fast

      Show the date beside each figure.

    Evidence

    • Your websiteHomepage: title tag just "OpenRouter"; H1 "The Unified Interface For Every Model"; "700T+ Monthly Tokens", "10M+ Global Users", "80+ Providers", "500+ Models"; "250k+ apps using OpenRouter with 4.2M+ users globally"; sign-up with Google, GitHub or MetaMask, then "Buy credits"; featured agents Replit, Hermes Agent and Cline.Source
    • Your blog"OpenRouter is Joining Stripe" (August 19, 2026): "We now process 10+ trillion tokens per day from 400+ AI models for a community of over 10 million developers and companies"; "at least 10x growth in inference volume every year"; OpenRouter keeps its name, product and roadmap; "Stripe brings a large customer network, data on how internet businesses grow".Source
    • HiringDirector, Growth Marketing: "More than 10 million users and 250k+ apps process 200T+ tokens on the platform every month, nearly all of it starting in self-serve." The role owns "signing up, buying their first credits, and growing their usage", paid media "with a paid budget of $5M+ for 2027", "the AEO/SEO roadmap for openrouter.ai", conversion optimization on "homepage, model pages, pricing, signup, credit purchase", and "incrementality and holdout testing".Source
  • 05

    The growth director owns "buying their first credits". Free users get 25+ free models at 50 requests a day, and the homepage shows credit purchases from $10. The point where a free user hits the daily limit is the strongest buying moment, and today it is just a limit.

    High impactLow effort First result in 1 month Director

    First action

    Count how many free users hit the limit each day.

    How to measure it

    First credit purchases from limit-hitters

    Now
    –
    Target
    +25%
    By
    60 days

    Steps

    1. 1

      Measure

      Daily limit hits and current conversion.

      Director, Growth Marketing 2 days

    2. 2

      Design the moment

      Estimate, button and model suggestions.

      Product Designer 3 days

    3. 3

      Build

      In chat, dashboard and API error docs.

      Software Engineer, Money Platform 1 week

    4. 4

      Test the bonus

      Bonus vs no bonus.

      Director, Growth Marketing 4 weeks

    5. 5

      Email follow-up

      Next day, for those who did not buy.

      Director, Growth Marketing 2 days

    6. 6

      Read out

      Conversion and 30-day spend.

      Director, Growth Marketing 6 weeks

    First result
    1 month
    Cost
    Engineering time and bonus credits
    Owner
    Director, Growth Marketing

    Smallest test

    Show the estimate to half of limit-hitters for two weeks.

    Why it works

    When a free user hits the limit, show what $10 buys on the models they used (estimated requests), a one-click credit purchase, and the paid models they could switch to. Test a small first-purchase bonus against no bonus.

    A user who hits a limit has already proven the product works for them; the offer lands at peak intent.

    Risks and how to handle them

    • Abuse of bonus credits

      Limit the bonus to verified first purchases.

    Evidence

    • HiringDirector, Growth Marketing: "More than 10 million users and 250k+ apps process 200T+ tokens on the platform every month, nearly all of it starting in self-serve." The role owns "signing up, buying their first credits, and growing their usage", paid media "with a paid budget of $5M+ for 2027", "the AEO/SEO roadmap for openrouter.ai", conversion optimization on "homepage, model pages, pricing, signup, credit purchase", and "incrementality and holdout testing".Source
    • Your websitePricing page: Free (25+ free models, "50 requests/day"), Standard (5.5% fee on credit purchases), Business (8%, adds EU and US in-region routing, advanced notifications, Workload Identity Federation and a 1,000-workspace limit) and Enterprise (fee discounts, SSO/SAML, SLAs, invoicing); BYOK includes "$25,000 of list price inference / month with no fees, 5% fee after"; failed attempts are not charged.Source
    • Your websiteHomepage: title tag just "OpenRouter"; H1 "The Unified Interface For Every Model"; "700T+ Monthly Tokens", "10M+ Global Users", "80+ Providers", "500+ Models"; "250k+ apps using OpenRouter with 4.2M+ users globally"; sign-up with Google, GitHub or MetaMask, then "Buy credits"; featured agents Replit, Hermes Agent and Cline.Source
  • 06

    Business charges 8% on credit purchases against 5.5% on Standard, in return for EU and US in-region routing, advanced notifications, Workload Identity Federation and up to 1,000 workspaces. A visitor sees a higher fee first and has to work out who should pay it.

    Medium impactLow effort First result in 3 weeks Director

    First action

    Write the three questions and the fee math.

    How to measure it

    Business upgrades

    Now
    –
    Target
    +20%
    By
    60 days

    Steps

    1. 1

      Agree the logic

      Questions and outcomes.

      Director, Growth Marketing 1 day

    2. 2

      Add the fee math

      Difference at the visitor's monthly spend.

      Web Engineer 2 days

    3. 3

      Design

      Above the comparison table.

      Product Designer 1 day

    4. 4

      Ship

      Pricing and Business pages.

      Web Engineer 1 day

    5. 5

      Test

      Against the current page.

      Director, Growth Marketing 3 weeks

    6. 6

      Route Enterprise answers

      To a demo form.

      Business Development Representative, Enterprise 1 day

    First result
    3 weeks
    Cost
    Engineering time
    Owner
    Director, Growth Marketing

    Smallest test

    Add a static three-question table first.

    Why it works

    Add a "Which plan?" block above the table: three questions (Do you need data to stay in the EU or US? More than five workspaces? Budget alerts in Slack?) that point to Standard, Business or Enterprise, with the monthly fee difference at the visitor's spend.

    Clear plan fit lifts upgrades and sends real enterprise buyers to sales instead of a self-serve plan that does not fit.

    Risks and how to handle them

    • Fewer Standard signups

      Keep Standard the default answer.

    Evidence

    • Your websitePricing page: Free (25+ free models, "50 requests/day"), Standard (5.5% fee on credit purchases), Business (8%, adds EU and US in-region routing, advanced notifications, Workload Identity Federation and a 1,000-workspace limit) and Enterprise (fee discounts, SSO/SAML, SLAs, invoicing); BYOK includes "$25,000 of list price inference / month with no fees, 5% fee after"; failed attempts are not charged.Source
    • Your websiteBusiness page: "Run inference in a region you choose" (EU or US, same key and account), "8% platform fee on credit purchases, no monthly minimum, no contract", a DPA through the Terms of Service with a signed copy on request; there is no APAC region option.Source
    • HiringDirector, Growth Marketing: "More than 10 million users and 250k+ apps process 200T+ tokens on the platform every month, nearly all of it starting in self-serve." The role owns "signing up, buying their first credits, and growing their usage", paid media "with a paid budget of $5M+ for 2027", "the AEO/SEO roadmap for openrouter.ai", conversion optimization on "homepage, model pages, pricing, signup, credit purchase", and "incrementality and holdout testing".Source
  • 07

    The homepage title tag is just "OpenRouter", and the compare pages, the biggest page set on the site, are titled "<model> vs <model> - AI Model Comparison" with no brand, while model pages end in "| OpenRouter". The growth role owns AEO/SEO for openrouter.ai.

    Medium impactLow effort First result in 1 month Director

    First action

    Change the homepage title tag.

    How to measure it

    Search click-through on changed templates

    Now
    –
    Target
    +10%
    By
    60 days

    Steps

    1. 1

      Rewrite homepage title

      Product and model count.

      Director, Growth Marketing 1 hour

    2. 2

      Fix compare template

      Add the brand.

      Web Engineer 1 hour

    3. 3

      Audit templates

      Collections, providers, benchmarks.

      Director, Growth Marketing 1 day

    4. 4

      Fix gaps

      Titles and descriptions.

      Web Engineer 1 day

    5. 5

      Annotate

      Mark the date in analytics.

      Director, Growth Marketing 1 hour

    6. 6

      Measure

      Search Console by template.

      Director, Growth Marketing 4 weeks

    First result
    1 month
    Cost
    $0
    Owner
    Director, Growth Marketing

    Smallest test

    Change the homepage title today.

    Why it works

    Change the homepage title to describe the product (for example "OpenRouter: one API for 500+ AI models"), add "| OpenRouter" to compare page titles, and check the other templates (collections, providers, benchmarks) for the same gap.

    Clear titles lift click-through from search and tell AI assistants which company the page belongs to.

    Risks and how to handle them

    • Brand searches unaffected

      Measure non-brand queries separately.

    Evidence

    • Your websiteHomepage: title tag just "OpenRouter"; H1 "The Unified Interface For Every Model"; "700T+ Monthly Tokens", "10M+ Global Users", "80+ Providers", "500+ Models"; "250k+ apps using OpenRouter with 4.2M+ users globally"; sign-up with Google, GitHub or MetaMask, then "Buy credits"; featured agents Replit, Hermes Agent and Cline.Source
    • Your websitePage titles: model pages end in "| OpenRouter" (e.g. "Claude Opus 5.5 - API Pricing & Benchmarks | OpenRouter"), but the 3,055 compare pages do not (e.g. "DeepSeek V4 Pro 0423 vs DeepSeek V4 Pro 0813 - AI Model Comparison").Source
    • HiringDirector, Growth Marketing: "More than 10 million users and 250k+ apps process 200T+ tokens on the platform every month, nearly all of it starting in self-serve." The role owns "signing up, buying their first credits, and growing their usage", paid media "with a paid budget of $5M+ for 2027", "the AEO/SEO roadmap for openrouter.ai", conversion optimization on "homepage, model pages, pricing, signup, credit purchase", and "incrementality and holdout testing".Source

This quarter 11

The core plays for the next three months.

  • 08

    Nearly all usage starts in self-serve, and the new enterprise BDR post values "product adoption signals". Teams at large companies already buying credits with a company email are the warmest enterprise accounts OpenRouter has.

    High impactMedium effort First result in 1 quarter Business Development Represen…

    First action

    Pull the top 50 company domains by self-serve spend.

    How to measure it

    Product-sourced enterprise opportunities

    Now
    –
    Target
    30 a quarter
    By
    2 quarters

    Steps

    1. 1

      Define signals

      Spend growth, members, keys, Business features.

      Data team 1 week

    2. 2

      Build the list

      Weekly, by company domain.

      Data team 2 weeks

    3. 3

      Write the opener

      What they run and what Enterprise adds.

      Product Marketing Manager, Enterprise 1 week

    4. 4

      Work the list

      Calls and email.

      Business Development Representative, Enterprise weekly

    5. 5

      Hand off

      To Strategic AEs with context.

      Strategic Account Executive ongoing

    6. 6

      Measure

      Conversion by signal.

      Sales Operations monthly

    First result
    1 quarter
    Cost
    Staff time only
    Owner
    Business Development Representative, Enterprise

    Smallest test

    Call the top 20 accounts this month.

    Why it works

    Each week, flag organizations by company domain with rising spend, several members or keys, or use of Business features, and give BDRs the account with its models, spend trend and features used. Outreach opens with what the team already runs and offers SSO, invoicing and an SLA.

    Selling to a team that already uses the product is faster and cheaper than cold outbound.

    Risks and how to handle them

    • Users feel watched

      Use only account-level data the terms allow and lead with help, not usage details.

    Evidence

    • HiringEnterprise sales hiring: a Business Development Representative, Enterprise (posted October 8, 2026) to "build qualified pipeline through outbound prospecting", with "Experience using product adoption signals to identify enterprise opportunities" as a plus; a Strategic Account Executive for "multi-stakeholder deals" across "CTO, CIO, CFO, and AI leadership" and "seven-figure deals"; plus Account Executives, Solutions Architect and Customer Engineers.Source
    • HiringDirector, Growth Marketing: "More than 10 million users and 250k+ apps process 200T+ tokens on the platform every month, nearly all of it starting in self-serve." The role owns "signing up, buying their first credits, and growing their usage", paid media "with a paid budget of $5M+ for 2027", "the AEO/SEO roadmap for openrouter.ai", conversion optimization on "homepage, model pages, pricing, signup, credit purchase", and "incrementality and holdout testing".Source
    • Your websitePricing page: Free (25+ free models, "50 requests/day"), Standard (5.5% fee on credit purchases), Business (8%, adds EU and US in-region routing, advanced notifications, Workload Identity Federation and a 1,000-workspace limit) and Enterprise (fee discounts, SSO/SAML, SLAs, invoicing); BYOK includes "$25,000 of list price inference / month with no fees, 5% fee after"; failed attempts are not charged.Source
  • 09

    The PMM role calls for advocacy, references and "a customer advisory board or design partner program". With only two customer stories published, each board member is also a candidate story.

    High impactMedium effort First result in 1 quarter Product Marketing Manager

    First action

    List ten customers to invite.

    How to measure it

    Customer stories live

    Now
    –
    Target
    8
    By
    2 quarters

    Steps

    1. 1

      Pick members

      Ten customers, mixed sizes.

      Product Marketing Manager, Enterprise 2 weeks

    2. 2

      Design the program

      Benefits and asks.

      Product Marketing Manager, Enterprise 1 week

    3. 3

      First meeting

      Roadmap and feedback.

      Product Marketing Manager, Enterprise 1 month

    4. 4

      Collect stories

      Interviews with each member.

      Product Marketing Manager, Enterprise ongoing

    5. 5

      Feed product

      Notes to product leads.

      Product Marketing Manager, Enterprise quarterly

    6. 6

      Track

      Stories, references and retention.

      Product Marketing Manager, Enterprise quarterly

    First result
    1 quarter
    Cost
    Events and travel
    Owner
    Product Marketing Manager, Enterprise

    Smallest test

    Invite three customers to a first call.

    Why it works

    Invite ten customers across the logo list and fast-growing self-serve teams to a quarterly board. Give them early access, roadmap sessions and a named engineer; ask each for a story, a reference call or a quote within two quarters.

    Advisory boards keep big customers close and produce the proof sales needs.

    Risks and how to handle them

    • Members expect roadmap control

      Share priorities, not promises.

    Evidence

    • HiringProduct Marketing Manager, Enterprise (posted September 29, 2026): owns the enterprise journey "from demo request to signed contract to expansion"; will "build the sales narrative and customer proof that help our biggest deals close", "Build and maintain competitor comparisons", "run win/loss analysis", and "Stand up and scale advocacy, case studies, references, and a customer advisory board or design partner program".Source
    • Your websiteEnterprise page (checked October 22, 2026): H1 "Your inference control plane"; a strip of eight customer logos (Lovable, Replit, Amplitude, Nvidia, Zoom, WTW, NIST, AMD) but no customer story, quote or named result anywhere on the page and no link to a case study; the only cost figure is an illustrative "Example: blended cost per 1M tokens -27%"; calls to action "Book a demo" and "Talk to us".Source
    • Your websiteCustomer proof on the site: two customer stories, /customers/descript ("1 -> 13 Models in production", "1+ week -> 1-2 hours To evaluate a promising new model") and /customers/paxhistoria ("35k" daily players, "4.8M+" monthly requests); the blog's case-study section adds Rasp, an agent used by OpenRouter's own sales team. The Descript story is published at three URLs, each with its own canonical: /customers/descript, /blog/case-studies/descript-case-study/ and /blog/insights/descript-model-evaluation-queue/.Source
  • 10

    OpenRouter Connect is a new channel for platforms to offer inference inside their products, and its first PM is choosing partner segments. There is no Connect page on the site, but the "Works with OpenRouter" directory already lists 53 tools that integrate.

    High impactMedium effort First result in 1 quarter Product Manager

    First action

    Score the 53 directory tools by fit.

    How to measure it

    Design partners

    Now
    –
    Target
    10
    By
    1 quarter

    Steps

    1. 1

      Write the page

      Value, controls, application.

      Product Manager, OpenRouter Connect 2 weeks

    2. 2

      Score tools

      Directory and app rankings.

      Product Manager, OpenRouter Connect 1 week

    3. 3

      Invite ten

      Personal outreach.

      Senior Manager, Product Partnerships 2 weeks

    4. 4

      Scope

      Needs per partner.

      Product Manager, OpenRouter Connect 1 month

    5. 5

      Set terms

      Revenue share options.

      Senior Manager, Product Partnerships 1 month

    6. 6

      Report

      Activation per partner.

      Product Manager, OpenRouter Connect monthly

    First result
    1 quarter
    Cost
    Staff time only
    Owner
    Product Manager, OpenRouter Connect

    Smallest test

    Interview five directory tools about Connect.

    Why it works

    Publish a Connect page explaining the model (platforms offer the catalog to their users with per-user controls and metering) with a design-partner application. Score the 53 listed tools and the top public apps by user count and fit, and invite the best ten.

    Tools that already integrate know the value and are the fastest design partners.

    Risks and how to handle them

    • Partners want custom work

      Keep to reusable building blocks, as the PM role requires.

    Evidence

    • HiringProduct Manager, OpenRouter Connect (the first PM for it): "Connect opens a new growth channel by letting partners offer multi-model inference to their customers inside their own products. We're building a platform for platforms"; the PM will "Determine and prioritize the target partner segments" and "Work directly with design partners".Source
    • Your websiteSitemaps (read October 22, 2026): 1,196 model pages, 3,055 model-vs-model compare pages, 527 docs pages (including model migration guides such as Claude 5.5 and GPT-6), 159 blog URLs, 16 collections, 9 rankings and benchmark pages, a /providers/apply form and 53 "Works with OpenRouter" pages; no page comparing OpenRouter with other AI gateways, no guide for moving from another gateway, and no OpenRouter Connect page.Source
    • Your websiteApp & Agent Rankings list the "Largest public apps and agents opting into usage tracking on OpenRouter", led by Hermes Agent at "53.1T" tokens, then Claude Code; the "Works with OpenRouter" directory has 53 tools.Source
  • 11

    Inbound provider interest "outpaces our capacity to action it", and the partner role checks metadata and pricing by hand. The /providers/apply form exists, but each provider still needs a person.

    Medium impactMedium effort First result in 1 quarter Partner Development Manager

    First action

    Write the provider checklist.

    How to measure it

    Median days to go live

    Now
    –
    Target
    -50%
    By
    2 quarters

    Steps

    1. 1

      Write the checklist

      From recent onboardings.

      Partner Development Manager, AI Ecosystem 1 week

    2. 2

      Build the test

      Conformance runner.

      Engineering Manager, Provider Ecosystem 1 month

    3. 3

      Status page

      Per application.

      Engineering Manager, Provider Ecosystem 2 weeks

    4. 4

      Triage rules

      Who gets a person.

      Partner Development Manager, AI Ecosystem 1 week

    5. 5

      Launch

      Announce to the inbound queue.

      Partner Development Manager, AI Ecosystem 1 week

    6. 6

      Measure

      Days to live and drop-off.

      Partner Development Manager, AI Ecosystem monthly

    First result
    1 quarter
    Cost
    Engineering time
    Owner
    Partner Development Manager, AI Ecosystem

    Smallest test

    Run five applicants through the checklist by hand.

    Build the monitor. A live page fed by the public registries and re-checked every day. Usually live within days.

    Book a working session

    Why it works

    Add a public provider checklist (API compatibility, metadata, pricing format, uptime test), an automated conformance test providers can run themselves, and a status page for each application. People handle only providers that pass.

    Automating the common steps lets a small team add many more providers, which is OpenRouter's supply advantage.

    Risks and how to handle them

    • Low-quality providers

      Keep uptime and accuracy gates before listing.

    Evidence

    • HiringPartnerships hiring: a Partner Development Manager, AI Ecosystem because "inbound interest ... outpaces our capacity to action it", who will "Source the long tail" of model developers and GPU providers; a Senior Manager, Product Partnerships to shape "revenue share, usage-based billing, and per-seat models" and build "playbooks, and templates so deal ten costs a fraction of deal one".Source
    • Your websiteSitemaps (read October 22, 2026): 1,196 model pages, 3,055 model-vs-model compare pages, 527 docs pages (including model migration guides such as Claude 5.5 and GPT-6), 159 blog URLs, 16 collections, 9 rankings and benchmark pages, a /providers/apply form and 53 "Works with OpenRouter" pages; no page comparing OpenRouter with other AI gateways, no guide for moving from another gateway, and no OpenRouter Connect page.Source
  • 12

    The Community Lead post asks for "the event kit that lets a community member in any city run their own" for a Discord of 50,000+ members.

    Medium impactMedium effort First result in 1 quarter Community Lead

    First action

    Ask the Discord which cities want an event.

    How to measure it

    Events held

    Now
    –
    Target
    20
    By
    2 quarters

    Steps

    1. 1

      Ask the community

      Cities and host volunteers.

      Community Lead, Developer Relations 1 week

    2. 2

      Build the kit

      Format, slides, demo.

      Community Lead, Developer Relations 3 weeks

    3. 3

      Budget

      Credits and food.

      Community Lead, Developer Relations 1 week

    4. 4

      Pick hosts

      Ten cities.

      Community Lead, Developer Relations 2 weeks

    5. 5

      List events

      One page.

      Community Lead, Developer Relations 1 week

    6. 6

      Measure

      Attendees and signups.

      Community Lead, Developer Relations monthly

    First result
    1 quarter
    Cost
    Event budget
    Owner
    Community Lead, Developer Relations

    Smallest test

    Support one member-run event.

    Why it works

    Build the kit: format, slides, demo, a credits budget, swag and a sign-up page. Use Discord to find hosts by city, start with ten cities, and list events on one page.

    Local meetups led by members spread trust far faster than company events.

    Risks and how to handle them

    • Uneven event quality

      Host checklist and a short call before each event.

    Evidence

    • HiringCommunity Lead, Developer Relations: the Discord has "Over 50,000 members, hundreds of threads, and thousands of chats every week". The work should be "1:n": "Instead of hosting one meetup, you create the event kit that lets a community member in any city run their own"; recruit a mod team; "Own the loop from community to product ... and close the loop back to the community when their input ships."Source
    • Public channelsDeveloper reach: @openrouter/ai-sdk-provider has about 1.95 million weekly npm downloads and @openrouter/sdk about 528,000; the YouTube channel has about 2,900 subscribers and 25 videos.Source
  • 13

    The Community Lead owns closing the loop "back to the community when their input ships". The docs changelog lists changes, but not which ones came from community requests.

    Medium impactLow effort First result in 1 month Community Lead

    First action

    Tag last month's shipped items that came from Discord.

    How to measure it

    Thread replies per post

    Now
    –
    Target
    200
    By
    2 quarters

    Steps

    1. 1

      Tag requests

      In the feedback tracker.

      Community Lead, Developer Relations 1 week

    2. 2

      Match shipped work

      From the changelog.

      Community Lead, Developer Relations monthly

    3. 3

      Write the post

      Short and specific.

      Community Lead, Developer Relations monthly

    4. 4

      Credit members

      With permission.

      Community Lead, Developer Relations monthly

    5. 5

      Share

      Discord, X and blog.

      Community Lead, Developer Relations monthly

    6. 6

      Measure

      Replies and new requests.

      Community Lead, Developer Relations monthly

    First result
    1 month
    Cost
    Staff time only
    Owner
    Community Lead, Developer Relations

    Smallest test

    Publish the first post this month.

    Why it works

    Tag community requests, and each month publish a short post and Discord thread naming what shipped from them, crediting the members who asked (with permission).

    Seeing their requests ship makes members advocates and keeps feedback coming.

    Risks and how to handle them

    • Too few items

      Combine months until there is enough.

    Evidence

    • HiringCommunity Lead, Developer Relations: the Discord has "Over 50,000 members, hundreds of threads, and thousands of chats every week". The work should be "1:n": "Instead of hosting one meetup, you create the event kit that lets a community member in any city run their own"; recruit a mod team; "Own the loop from community to product ... and close the loop back to the community when their input ships."Source
    • Your websiteSitemaps (read October 22, 2026): 1,196 model pages, 3,055 model-vs-model compare pages, 527 docs pages (including model migration guides such as Claude 5.5 and GPT-6), 159 blog URLs, 16 collections, 9 rankings and benchmark pages, a /providers/apply form and 53 "Works with OpenRouter" pages; no page comparing OpenRouter with other AI gateways, no guide for moving from another gateway, and no OpenRouter Connect page.Source
  • 14

    The GRC Manager will curate a SafeBase knowledge base "so sales can self-serve most customer security questionnaires". The enterprise page talks about SOC 2 and DPAs but does not send security reviewers to the trust center.

    High impactMedium effort First result in 1 quarter GRC Manager

    First action

    List the ten questions security teams ask most.

    How to measure it

    Median security review time

    Now
    –
    Target
    -30%
    By
    2 quarters

    Steps

    1. 1

      List common questions

      From recent deals.

      GRC Manager 1 week

    2. 2

      Fill the knowledge base

      Answers and documents.

      GRC Manager 1 month

    3. 3

      Link from enterprise page

      For security teams.

      Product Marketing Manager, Enterprise 1 day

    4. 4

      Pre-fill a questionnaire

      A standard one.

      GRC Manager 2 weeks

    5. 5

      Train sellers

      Send the link first.

      Strategic Account Executive 1 day

    6. 6

      Measure

      Review time per deal.

      Sales Operations monthly

    First result
    1 quarter
    Cost
    Staff time only
    Owner
    GRC Manager

    Smallest test

    Add the trust center link to the enterprise page.

    Why it works

    Add a "For security teams" block on the enterprise page linking the trust center, the DPA, data retention docs and a pre-filled standard questionnaire. Track how many deals start a security review with the documents already downloaded.

    Security review is often the longest step in an enterprise deal, and self-serve documents shorten it.

    Risks and how to handle them

    • Documents go stale

      Quarterly review owned by GRC.

    Evidence

    • HiringGRC Manager (posted September 28, 2026): will "Curate the SafeBase knowledge base so sales can self-serve most customer security questionnaires" and keep audit readiness "across SOC 2, HIPAA, PCI DSS, GDPR, CCPA, BIPA, and the EU AI Act".Source
    • Your websiteEnterprise page (checked October 22, 2026): H1 "Your inference control plane"; a strip of eight customer logos (Lovable, Replit, Amplitude, Nvidia, Zoom, WTW, NIST, AMD) but no customer story, quote or named result anywhere on the page and no link to a case study; the only cost figure is an illustrative "Example: blended cost per 1M tokens -27%"; calls to action "Book a demo" and "Talk to us".Source
    • Your websiteBusiness page: "Run inference in a region you choose" (EU or US, same key and account), "8% platform fee on credit purchases, no monthly minimum, no contract", a DPA through the Terms of Service with a signed copy on request; there is no APAC region option.Source
  • 15

    OpenRouter is hiring a foundational Customer Engineer for its "expanding APAC footprint", who will assess data residency in discovery. The Business plan offers EU and US routing only, and the site has no APAC page.

    Medium impactMedium effort First result in 2 quarters Customer Engineer

    First action

    List APAC customers and their main questions.

    How to measure it

    APAC demos

    Now
    –
    Target
    20 a quarter
    By
    2 quarters

    Steps

    1. 1

      Gather questions

      From current APAC customers.

      Customer Engineer, APAC 2 weeks

    2. 2

      Write the page

      Latency, providers, data, support.

      Customer Engineer, APAC 2 weeks

    3. 3

      Pick two markets

      By current usage.

      Director, Growth Marketing 1 week

    4. 4

      Run the campaign

      Developer channels.

      Director, Growth Marketing 1 month

    5. 5

      Track

      Signups and demos by country.

      Director, Growth Marketing monthly

    6. 6

      Decide

      Regional routing case.

      Customer Engineer, APAC 1 quarter

    First result
    2 quarters
    Cost
    Campaign budget
    Owner
    Customer Engineer, APAC

    Smallest test

    Publish the page and measure one month of traffic.

    Why it works

    Publish an APAC page covering latency, providers and models with endpoints in the region, data handling, payment options and support hours. Run a small developer campaign in two markets and track signups and enterprise demos by country.

    A page and a test show where APAC demand is before more hiring and any regional routing work.

    Risks and how to handle them

    • Promising features not built

      Describe only what is live today.

    Evidence

    • HiringCustomer Engineer, APAC (posted October 9, 2026): a "foundational" role for "OpenRouter's expanding APAC footprint", assessing "data residency, and compliance" in discovery and covering support and success "as we build out a regional team".Source
    • Your websiteBusiness page: "Run inference in a region you choose" (EU or US, same key and account), "8% platform fee on credit purchases, no monthly minimum, no contract", a DPA through the Terms of Service with a signed copy on request; there is no APAC region option.Source
    • Your websitePricing page: Free (25+ free models, "50 requests/day"), Standard (5.5% fee on credit purchases), Business (8%, adds EU and US in-region routing, advanced notifications, Workload Identity Federation and a 1,000-workspace limit) and Enterprise (fee discounts, SSO/SAML, SLAs, invoicing); BYOK includes "$25,000 of list price inference / month with no fees, 5% fee after"; failed attempts are not charged.Source
  • 16

    The growth director gets a paid budget of $5M+ for 2027 and is asked for "incrementality and holdout testing". Developer channels (newsletters, podcasts, creators) are hard to measure by clicks.

    High impactHigh effort First result in 2 quarters Director

    First action

    Pick three channels for the first holdout tests.

    How to measure it

    Incremental CAC

    Now
    –
    Target
    Below payback target
    By
    2 quarters

    Steps

    1. 1

      Set the metric

      First credit purchase and 90-day spend.

      Director, Growth Marketing 1 week

    2. 2

      Design tests

      Geo or audience holdouts.

      Director, Growth Marketing 2 weeks

    3. 3

      Run three channels

      Newsletter, podcast, search.

      Director, Growth Marketing 1 quarter

    4. 4

      Read out

      Lift per channel.

      Director, Growth Marketing 1 month

    5. 5

      Scale winners

      Into the 2027 plan.

      Director, Growth Marketing quarterly

    6. 6

      Repeat

      For new channels.

      Director, Growth Marketing quarterly

    First result
    2 quarters
    Cost
    Test budget
    Owner
    Director, Growth Marketing

    Smallest test

    Run one newsletter holdout test.

    Why it works

    Before scaling, test each channel with geo or audience holdouts and promo-free tracking (signup survey plus credit-purchase data). Scale only channels that show lift in first credit purchases.

    Holdout tests stop the budget going to channels that only take credit for signups that would have happened anyway.

    Risks and how to handle them

    • Small tests are noisy

      Run long enough for a clear read.

    Evidence

    • HiringDirector, Growth Marketing: "More than 10 million users and 250k+ apps process 200T+ tokens on the platform every month, nearly all of it starting in self-serve." The role owns "signing up, buying their first credits, and growing their usage", paid media "with a paid budget of $5M+ for 2027", "the AEO/SEO roadmap for openrouter.ai", conversion optimization on "homepage, model pages, pricing, signup, credit purchase", and "incrementality and holdout testing".Source
    • Public channelsDeveloper reach: @openrouter/ai-sdk-provider has about 1.95 million weekly npm downloads and @openrouter/sdk about 528,000; the YouTube channel has about 2,900 subscribers and 25 videos.Source
  • 17

    OpenRouter has 1,196 model pages titled for "API Pricing & Benchmarks", and the growth role owns answer-engine visibility. When a developer asks an assistant for a model's price or provider options, OpenRouter's page should be the cited source.

    High impactMedium effort First result in 2 quarters Director

    First action

    Check which sources assistants cite for 20 top models.

    How to measure it

    Share of cited answers

    Now
    –
    Target
    +20 points
    By
    2 quarters

    Steps

    1. 1

      Baseline

      50 models across assistants.

      Director, Growth Marketing 1 week

    2. 2

      Structured data

      Price, context, providers, date.

      Web Engineer 1 month

    3. 3

      Add FAQs

      Three per model.

      Content team 1 month

    4. 4

      Monitor

      Weekly citation checks.

      Director, Growth Marketing weekly

    5. 5

      Day-0 pages

      Live at each launch.

      Director, Growth Marketing ongoing

    6. 6

      Report

      Signups from model pages.

      Director, Growth Marketing monthly

    First result
    2 quarters
    Cost
    Engineering time
    Owner
    Director, Growth Marketing

    Smallest test

    Add structured data to 20 model pages and compare.

    Why it works

    Add consistent structured data (price per million tokens, context, providers, date updated) and a short FAQ to each model page, and check weekly which sources assistants cite for the top 50 models.

    Model searches are the moment a developer picks a route to a model; being the cited answer brings signups.

    Risks and how to handle them

    • Assistants change sources often

      Track trends, not single answers.

    Evidence

    • HiringDirector, Growth Marketing: "More than 10 million users and 250k+ apps process 200T+ tokens on the platform every month, nearly all of it starting in self-serve." The role owns "signing up, buying their first credits, and growing their usage", paid media "with a paid budget of $5M+ for 2027", "the AEO/SEO roadmap for openrouter.ai", conversion optimization on "homepage, model pages, pricing, signup, credit purchase", and "incrementality and holdout testing".Source
    • Your websiteSitemaps (read October 22, 2026): 1,196 model pages, 3,055 model-vs-model compare pages, 527 docs pages (including model migration guides such as Claude 5.5 and GPT-6), 159 blog URLs, 16 collections, 9 rankings and benchmark pages, a /providers/apply form and 53 "Works with OpenRouter" pages; no page comparing OpenRouter with other AI gateways, no guide for moving from another gateway, and no OpenRouter Connect page.Source
    • Your websitePage titles: model pages end in "| OpenRouter" (e.g. "Claude Opus 5.5 - API Pricing & Benchmarks | OpenRouter"), but the 3,055 compare pages do not (e.g. "DeepSeek V4 Pro 0423 vs DeepSeek V4 Pro 0813 - AI Model Comparison").Source
  • 18

    Ori is meant for "massive enterprise adoption of OpenRouter beyond the engineering organization". The site's Ori pages are written for developers and coding harnesses.

    Medium impactMedium effort First result in 2 quarters Software Engineer

    First action

    Pick one customer and one team.

    How to measure it

    Non-engineering Ori users at pilot customers

    Now
    –
    Target
    50
    By
    2 quarters

    Steps

    1. 1

      Pick the customer

      Strong relationship.

      Strategic Account Executive 2 weeks

    2. 2

      Pick the team

      Clear repetitive work.

      Software Engineer, Ori 1 week

    3. 3

      Set guardrails

      Budgets and models.

      Solutions Architect 1 week

    4. 4

      Run

      Six weeks.

      Software Engineer, Ori 6 weeks

    5. 5

      Measure

      Hours saved and cost.

      Software Engineer, Ori 1 week

    6. 6

      Publish

      Customer story.

      Product Marketing Manager, Enterprise 1 month

    First result
    2 quarters
    Cost
    Engineering support
    Owner
    Software Engineer, Ori

    Smallest test

    Run Ori with one team for two weeks.

    Why it works

    Run Ori with one non-engineering team (for example finance or sales operations) at an existing customer, with org guardrails and budgets, and publish the result as a story. The sales agent story about OpenRouter's own team is the template.

    One outside success outside engineering shows enterprise buyers a second use and widens the account.

    Risks and how to handle them

    • Pilot stalls

      Pick a team with a sponsor.

    Evidence

    • HiringSoftware Engineer, Ori (posted October 6, 2026): Ori is "building the client side products for massive enterprise adoption of OpenRouter beyond the engineering organization", across the CLI and its harness integrations; the site has Ori harness pages for 16 coding agents.Source
    • Your websiteCustomer proof on the site: two customer stories, /customers/descript ("1 -> 13 Models in production", "1+ week -> 1-2 hours To evaluate a promising new model") and /customers/paxhistoria ("35k" daily players, "4.8M+" monthly requests); the blog's case-study section adds Rasp, an agent used by OpenRouter's own sales team. The Descript story is published at three URLs, each with its own canonical: /customers/descript, /blog/case-studies/descript-case-study/ and /blog/insights/descript-model-evaluation-queue/.Source

Big bets 3

Larger plays that can change the trajectory.

  • 19

    OpenRouter joined Stripe in August 2026, and its announcement says Stripe "brings a large customer network". Many businesses that bill through Stripe are adding AI features and need models, budgets and cost tracking.

    High impactHigh effort First result in 3 quarters Senior Manager

    First action

    List joint accounts already using both.

    How to measure it

    Stripe-sourced enterprise opportunities

    Now
    –
    Target
    50
    By
    3 quarters

    Steps

    1. 1

      Agree the plan

      Channels and rules.

      Senior Manager, Product Partnerships 1 month

    2. 2

      Build the guide

      AI features for Stripe users.

      Content team 1 month

    3. 3

      Joint accounts

      Shared list.

      Strategic Account Executive 1 month

    4. 4

      Launch

      Developer channels.

      Director, Growth Marketing 1 quarter

    5. 5

      Co-sell

      Joint calls.

      Strategic Account Executive ongoing

    6. 6

      Measure

      Source tracking.

      Sales Operations monthly

    First result
    3 quarters
    Cost
    Staff time
    Owner
    Senior Manager, Product Partnerships

    Smallest test

    Run one joint webinar.

    Why it works

    Agree a co-marketing and referral plan with Stripe: OpenRouter content and offers in Stripe's developer channels, a guide for Stripe users adding AI features, and joint enterprise accounts where the same buyer runs both. Measure signups and enterprise deals from Stripe sources.

    A trusted parent brand with millions of developer accounts is the cheapest new reach OpenRouter has.

    Risks and how to handle them

    • Neutrality concerns

      Keep routing decisions independent, as the announcement promises.

    Evidence

    • Your blog"OpenRouter is Joining Stripe" (August 19, 2026): "We now process 10+ trillion tokens per day from 400+ AI models for a community of over 10 million developers and companies"; "at least 10x growth in inference volume every year"; OpenRouter keeps its name, product and roadmap; "Stripe brings a large customer network, data on how internet businesses grow".Source
    • HiringPartnerships hiring: a Partner Development Manager, AI Ecosystem because "inbound interest ... outpaces our capacity to action it", who will "Source the long tail" of model developers and GPU providers; a Senior Manager, Product Partnerships to shape "revenue share, usage-based billing, and per-seat models" and build "playbooks, and templates so deal ten costs a fraction of deal one".Source
    • HiringEnterprise sales hiring: a Business Development Representative, Enterprise (posted October 8, 2026) to "build qualified pipeline through outbound prospecting", with "Experience using product adoption signals to identify enterprise opportunities" as a plus; a Strategic Account Executive for "multi-stakeholder deals" across "CTO, CIO, CFO, and AI leadership" and "seven-figure deals"; plus Account Executives, Solutions Architect and Customer Engineers.Source
  • 20

    The State of AI 2025 study with a16z covered 100 trillion tokens, the data page says it is cited by MIT, NIST and a16z, and the Inferred newsletter now publishes weekly. Volume has grown at least tenfold a year since.

    High impactHigh effort First result in 3 quarters Director

    First action

    Pick the five charts for the first quarterly update.

    How to measure it

    Report downloads

    Now
    –
    Target
    50,000
    By
    1 year

    Steps

    1. 1

      Plan topics

      Five recurring charts.

      Director, Growth Marketing 2 weeks

    2. 2

      Analyze

      Aggregated usage only.

      Data team 1 month

    3. 3

      Privacy review

      No customer identifiable.

      GRC Manager 2 weeks

    4. 4

      Publish

      Report and summary.

      Director, Growth Marketing 1 month

    5. 5

      Launch

      Press, webinar, social.

      Director, Growth Marketing 1 month

    6. 6

      Use in sales

      Account-level talk tracks.

      Product Marketing Manager, Enterprise ongoing

    First result
    3 quarters
    Cost
    Analyst time
    Owner
    Director, Growth Marketing

    Smallest test

    Publish one quarterly chart this month.

    Why it works

    Plan the 2026 edition with a fixed release date, then publish short quarterly updates on model share, open vs closed, agents and regions, each with a press briefing and a webinar for enterprise buyers.

    Original data earns press, links and AI citations, and gives sales a reason to call every quarter.

    Risks and how to handle them

    • Data privacy

      Aggregate and suppress small groups.

    Evidence

    • Your websiteData: the "State of AI 2025" report with a16z (December 2025), "An Empirical 100 Trillion Token Study"; the /data page says the data is "Cited by MIT, NIST, a16z, and more" and "We don't sell prompt data"; a weekly newsletter, Inferred, with three editions (September 22, October 1 and October 8, 2026).Source
    • Your blog"OpenRouter is Joining Stripe" (August 19, 2026): "We now process 10+ trillion tokens per day from 400+ AI models for a community of over 10 million developers and companies"; "at least 10x growth in inference volume every year"; OpenRouter keeps its name, product and roadmap; "Stripe brings a large customer network, data on how internet businesses grow".Source
  • 21

    The App & Agent Rankings show which apps send the most tokens, led by Hermes Agent at 53.1T. Apps that already route large volumes and serve their own users are the natural first Connect partners.

    High impactHigh effort First result in 4 quarters Product Manager

    First action

    Rank the top 50 apps by fit.

    How to measure it

    Apps on Connect

    Now
    –
    Target
    20
    By
    4 quarters

    Steps

    1. 1

      Rank apps

      By tokens and fit.

      Product Manager, OpenRouter Connect 2 weeks

    2. 2

      Design terms

      Revenue share.

      Senior Manager, Product Partnerships 1 month

    3. 3

      Pitch top 20

      Personal outreach.

      Senior Manager, Product Partnerships 1 quarter

    4. 4

      Integrate

      Design partners first.

      Product Manager, OpenRouter Connect 2 quarters

    5. 5

      Support

      Named engineer.

      Solutions Architect ongoing

    6. 6

      Measure

      Usage and revenue.

      Product Manager, OpenRouter Connect monthly

    First result
    4 quarters
    Cost
    Partnership time
    Owner
    Product Manager, OpenRouter Connect

    Smallest test

    Pitch three top apps.

    Why it works

    Rank apps by tokens, user count and whether they resell inference to their users, and offer the top 20 Connect with revenue share, per-user controls and metering. Track usage moving into Connect.

    The best partners are already customers, and the rankings show who they are.

    Risks and how to handle them

    • Apps prefer full control

      Offer Connect as an option, not a requirement.

    Evidence

    • Your websiteApp & Agent Rankings list the "Largest public apps and agents opting into usage tracking on OpenRouter", led by Hermes Agent at "53.1T" tokens, then Claude Code; the "Works with OpenRouter" directory has 53 tools.Source
    • HiringProduct Manager, OpenRouter Connect (the first PM for it): "Connect opens a new growth channel by letting partners offer multi-model inference to their customers inside their own products. We're building a platform for platforms"; the PM will "Determine and prioritize the target partner segments" and "Work directly with design partners".Source
    • HiringPartnerships hiring: a Partner Development Manager, AI Ecosystem because "inbound interest ... outpaces our capacity to action it", who will "Source the long tail" of model developers and GPU providers; a Senior Manager, Product Partnerships to shape "revenue share, usage-based billing, and per-seat models" and build "playbooks, and templates so deal ten costs a fraction of deal one".Source

Wildcards 2

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

  • 22

    The enterprise page shows an example of a blended cost 27% lower when routed, and providers compete on price for the same model. A buyer cannot yet see what routing would save on their own usage.

    Medium impactMedium effort First result in 1 quarter Product Marketing Manager

    First action

    Agree the calculation with finance.

    How to measure it

    Demos from calculator

    Now
    –
    Target
    50 a month
    By
    2 quarters

    Steps

    1. 1

      Agree the method

      Live prices and assumptions.

      Product Marketing Manager, Enterprise 1 week

    2. 2

      Design

      Inputs and result.

      Product Designer 1 week

    3. 3

      Build

      Using price data.

      Web Engineer 1 month

    4. 4

      PDF export

      For finance teams.

      Web Engineer 1 week

    5. 5

      Launch

      Enterprise and pricing pages.

      Product Marketing Manager, Enterprise 1 week

    6. 6

      Measure

      Demos and deal size.

      Sales Operations monthly

    First result
    1 quarter
    Cost
    Engineering time
    Owner
    Product Marketing Manager, Enterprise

    Smallest test

    Run the math by hand for five prospects.

    Why it works

    Build a calculator: the visitor enters monthly tokens per model, and it shows cost direct vs routed with live provider prices, plus fallback uptime. Offer the result as a PDF for the buyer's finance team and a demo.

    Finance buyers act on their own numbers, and the calculator gives AEs a ready business case.

    Risks and how to handle them

    • Overstated savings

      Show assumptions and ranges.

    Evidence

    • Your websiteEnterprise page (checked October 22, 2026): H1 "Your inference control plane"; a strip of eight customer logos (Lovable, Replit, Amplitude, Nvidia, Zoom, WTW, NIST, AMD) but no customer story, quote or named result anywhere on the page and no link to a case study; the only cost figure is an illustrative "Example: blended cost per 1M tokens -27%"; calls to action "Book a demo" and "Talk to us".Source
    • Your websitePricing page: Free (25+ free models, "50 requests/day"), Standard (5.5% fee on credit purchases), Business (8%, adds EU and US in-region routing, advanced notifications, Workload Identity Federation and a 1,000-workspace limit) and Enterprise (fee discounts, SSO/SAML, SLAs, invoicing); BYOK includes "$25,000 of list price inference / month with no fees, 5% fee after"; failed attempts are not charged.Source
    • HiringEnterprise sales hiring: a Business Development Representative, Enterprise (posted October 8, 2026) to "build qualified pipeline through outbound prospecting", with "Experience using product adoption signals to identify enterprise opportunities" as a plus; a Strategic Account Executive for "multi-stakeholder deals" across "CTO, CIO, CFO, and AI leadership" and "seven-figure deals"; plus Account Executives, Solutions Architect and Customer Engineers.Source
  • 23

    The npm SDK provider has about 1.95 million weekly downloads and the Discord 50,000+ members, but YouTube has about 2,900 subscribers and 25 videos. The weekly Inferred newsletter already produces charts each week.

    Medium impactLow effort First result in 1 quarter Community Lead

    First action

    Record a video for the latest edition.

    How to measure it

    Subscribers

    Now
    –
    Target
    20,000
    By
    3 quarters

    Steps

    1. 1

      Pick format

      One chart, one minute.

      Community Lead, Developer Relations 1 week

    2. 2

      Record

      Weekly.

      Community Lead, Developer Relations weekly

    3. 3

      Guest episodes

      Monthly.

      Community Lead, Developer Relations monthly

    4. 4

      Distribute

      YouTube, X, LinkedIn.

      Community Lead, Developer Relations weekly

    5. 5

      Link back

      To newsletter and signup.

      Director, Growth Marketing weekly

    6. 6

      Measure

      Views and signups.

      Director, Growth Marketing monthly

    First result
    1 quarter
    Cost
    Production time
    Owner
    Community Lead, Developer Relations

    Smallest test

    Make three videos.

    Why it works

    Make one short video per newsletter edition (the main chart and what it means), plus a monthly longer video with a community guest. Post to YouTube, X and LinkedIn.

    The data is already made; video reaches developers who never open a newsletter.

    Risks and how to handle them

    • Time cost

      Keep videos simple.

    Evidence

    • Public channelsDeveloper reach: @openrouter/ai-sdk-provider has about 1.95 million weekly npm downloads and @openrouter/sdk about 528,000; the YouTube channel has about 2,900 subscribers and 25 videos.Source
    • Your websiteData: the "State of AI 2025" report with a16z (December 2025), "An Empirical 100 Trillion Token Study"; the /data page says the data is "Cited by MIT, NIST, a16z, and more" and "We don't sell prompt data"; a weekly newsletter, Inferred, with three editions (September 22, October 1 and October 8, 2026).Source
    • HiringCommunity Lead, Developer Relations: the Discord has "Over 50,000 members, hundreds of threads, and thousands of chats every week". The work should be "1:n": "Instead of hosting one meetup, you create the event kit that lets a community member in any city run their own"; recruit a mod team; "Own the loop from community to product ... and close the loop back to the community when their input ships."Source

03

How Programmable can help

Where an idea in this playbook is something Programmable builds or runs, it's listed here with what we'd do. Everything else is yours to run.

Have it built

ProgrammableAgency

1 of the 23 ideas are things Programmable builds and runs for its customers: pages, outreach, prospect lists, booking flows and monitors. Pick one, or send the whole list.

Monitors1A live page fed by the public registries and re-checked every day.

A 30-minute session with the engineer who would build it, not a sales call. You leave with a plan whether or not you hire us.

Book a working session

Usually live within days.

Tools for the rest

ProgrammableTalk

Fill the events in this plan

These ideas are events. Talk gets registrants for them and tells you who's coming and what their company does before the day.

Plan an event (opens programmable.talk in a new tab)
ProgrammableDigital

Show it, don't just describe it

These ideas need something to show: a demo, a video, an animation, graphics or a deck. Digital turns screenshots or the site into all of them, in the company's own brand.

Make a demo (opens programmable.digital in a new tab)
More example playbooks and reviews

More growth playbooks

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

  • ProsperSpark growth strategyAn Omaha-based consultancy that builds custom spreadsheets (Excel, VBA, Power Query, Power BI), automations (Make, Zapier, Power Automate)…22 growth ideas
  • SARELLY growth strategyA Mexican makeup and fashion-accessories brand co-founded by creator Anna Sarelly: waterproof mascaras, lips, palettes and kits, plus bags…22 growth ideas
  • Helium 10 growth strategySoftware for marketplace sellers: product and keyword research, listing optimization, ads automation, profits and operations tools, a…22 growth ideas
  • Tekmetric growth strategyCloud shop management software for auto repair shops, with digital inspections, estimates, parts, reporting, and add-ons for multi-shop…22 growth ideas
  • Pineo growth strategyA London HubSpot consultancy, Platinum partner for both HubSpot and Zapier. It offers fixed-price HubSpot onboarding, implementation…23 growth ideas
  • Hemut growth strategyAn AI operating system for trucking carriers: Command (dispatch, live tracking, documents, invoicing, settlements and IFTA, with the Hemi…23 growth ideas
  • Auctor growth strategyAn AI-native "system of action" for software implementations: it captures discovery calls, documents and decisions, keeps them traceable…23 growth ideas
  • Aurify growth strategyA consumer AI photo app: upload one selfie, pick a trending scene, and get yourself placed into cinematic, travel, luxury and fashion…20 growth ideas

See every playbook and website review

Independent research. Programmable wrote this playbook from public sources, as they stood on October 23, 2026, and cites them throughout. OpenRouter did not commission, review or endorse it, and Programmable is not affiliated with OpenRouter. 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 OpenRouter runs its business.

Something wrong, or do you represent OpenRouter? 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

    Enterprise proof, one Descript URL, gateway comparisons, one set of numbers, the free-limit moment, plan clarity and search titles.

  2. Days 31 to 6013 ideas

    Product-signal outbound, advisory board, Connect page, provider onboarding, meetups, community changelog, trust center, APAC, holdout tests, model-page AEO and an Ori pilot; start the calculator and videos.

  3. Days 61 to 903 ideas

    Commit to the big bets: Stripe distribution, recurring State of AI reports and Connect from the app rankings.

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 integration for every model. The most cited problem: model turnover every few weeks.

  • New models on day 0
  • Provider outages and rate limits
  • In their words: “one API”

Market

Competitive landscape

5 competitors compared: Portkey, LiteLLM, Vercel AI Gateway, Cloudflare AI Gateway and 1 more.

  • Gap to exploit: Published enterprise results for multi-model gateways
  • Portkey: No marketplace or usage data at OpenRouter's scale
  • LiteLLM: Teams run and maintain it themselves

Hiring signals

OpenRouter is building an enterprise and partner business on top of its self-serve scale.

  • Product Marketing Manager, Enterprise: Customer proof, competitor comparisons and an advisory board. Ideas 1, 3, 9.
  • Director, Growth Marketing: $5M+ paid budget, AEO/SEO and first credit purchases. Ideas 2, 5, 6, 7, 16, 17.
  • Business Development Representative, Enterprise: Outbound using product adoption signals. Idea 8.

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 firstNew accounts get 100,000 free enrichment credits with code BETA100K (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)
  • ProgrammableMediaAn AI note taker, interpreter and webinar host for your calls (opens programmable.media 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