Trigger.dev growth playbook

Publish /vs/vercel-workflow this week in the same format as /vs/inngest: a fair "choose Workflow SDK if / choose Trigger.dev if" summary, a limits table that quotes your docs (no maximum run duration, 14-day maximum queue wait, automatic payload offloading, Python via the build extension), and a short "long human-in-the-loop flows" example with a wait longer than two weeks. Add it to llms.txt, /for-ai-crawlers and the .md index, then send Vercel's docs team a polite note with links to the limits page so their table can be updated.

Comparison queries come from developers close to choosing, and AI assistants quote whichever page states the facts most plainly. A dated, sourced page of your own turns the most important rival comparison from their framing into yours, and is exactly the GEO work the Marketing Lead is being hired for.

H1 Trigger.dev vs Vercel Workflow SDK
Summary: choose Workflow SDK if your workflows live inside your Vercel app and you accept deterministic replay; choose Trigger.dev if you want plain async code on managed compute with no run timeout
Limits table: run duration (none) | queue wait (14 days max TTL) | payloads (auto-offloaded over 128KB) | Python (pythonExtension)
Example: human approval after 3 weeks
Migration: concepts mapped (use workflow / use step -> task / triggerAndWait)
Sources: links to /docs/limits and the Workflow SDK page

• Looks combative

  Keep it factual, give Vercel's strengths their due, and link their page.

• Limits change later

  Generate the table from the limits page so both stay in step.

• Competitor researchVercel's Workflow SDK docs publish "Workflow SDK vs trigger.dev" (with "Copy for LLM" and "View Markdown" buttons). Its table lists trigger.dev limits as "3 MB payload / 10 MB output; 14-day maximum run lifetime", says "the 14-day maximum run time means human-in-the-loop flows that wait longer than two weeks can't complete in one run", that large model contexts "need external storage", and that trigger.dev is "TypeScript / JavaScript only". The same docs compare the Workflow SDK with Temporal, Cloudflare Workflows, AWS Step Functions, AWS Bedrock AgentCore and Inngest.See the research Source
• Your docsLimits page: "Runs have no maximum duration." Under the heading "Maximum run TTL": "On Trigger.dev Cloud, all runs have an enforced maximum TTL of 14 days" and "TTL governs how long a run may sit queued waiting for a concurrency slot." On payloads: "Large payloads and outputs are offloaded to object storage automatically. You don't need to do anything to handle this in your tasks."Source
• Your docsThe Python build extension: "If you need to execute Python scripts in your Trigger.dev project, you can use the pythonExtension build extension via the @trigger.dev/python package." The homepage use-case tabs include "Run Python".Source
• Your websiteSitemap (245 URLs, checked October 22, 2026): 165 changelog entries, 29 blog posts, 13 customer stories, 13 launch-week pages, 6 product pages; comparison pages only for Temporal, Inngest, BullMQ and n8n, plus "Claude Managed Agents vs Trigger.dev" and an agent-frameworks comparison. The homepage tabs list use cases (AI agents, Media processing, Media generation, Human in the loop, Streaming, Run Python, Marketing, Browser automation, Semantic search, Email sequences) but none has its own page. There is a /for-ai-crawlers page and an llms.txt, and every page is available as Markdown.Source
• HiringMarketing Lead (posted September 16, 2026, remote CET to ET or London): "own growth marketing at Trigger.dev: influencer partnerships, GEO/SEO, and the bold, high-visibility campaigns that get us noticed"; "picks one or two and goes deep"; influencer work "using tools like Plug.dev"; "big swing" campaigns such as "an SF billboard, or premium merch sends to YC founders"; and "narrow, targeted paid spend (e.g. a small daily budget tied to a free tool)".Source

Change the heading to "Maximum queue wait (TTL): 14 days", open the section with "This does not limit how long a run executes", and add the same sentence beside "Runs have no maximum duration". Do the same in the payload section: lead with "offloaded automatically", then the 3MB and 10MB thresholds.

Docs are now read by the AI tools that recommend you (your own job posts say so), and a heading is what they quote. One wording change removes the misreading at the source.

• Breaks existing anchors

  Keep the old anchor id as an alias.

• Your docsLimits page: "Runs have no maximum duration." Under the heading "Maximum run TTL": "On Trigger.dev Cloud, all runs have an enforced maximum TTL of 14 days" and "TTL governs how long a run may sit queued waiting for a concurrency slot." On payloads: "Large payloads and outputs are offloaded to object storage automatically. You don't need to do anything to handle this in your tasks."Source
• Competitor researchVercel's Workflow SDK docs publish "Workflow SDK vs trigger.dev" (with "Copy for LLM" and "View Markdown" buttons). Its table lists trigger.dev limits as "3 MB payload / 10 MB output; 14-day maximum run lifetime", says "the 14-day maximum run time means human-in-the-loop flows that wait longer than two weeks can't complete in one run", that large model contexts "need external storage", and that trigger.dev is "TypeScript / JavaScript only". The same docs compare the Workflow SDK with Temporal, Cloudflare Workflows, AWS Step Functions, AWS Bedrock AgentCore and Inngest.See the research Source
• HiringEvery Trigger.dev job post: "We are a product-led growth company so everyone is expected to get involved in creating content like code examples, blog articles, videos and tweets" and "Great documentation matters more than ever, both for humans and for the AI tools that read it." The Senior Backend Engineer post says the Cloud serves "thousands of teams building AI apps and agents, handling hundreds of millions of executions per month"; a Backend Engineer (Security) role is also open.Source

Write five pages in the /vs/ format, each with a fair "choose X if" summary, a code-to-code example and a migration section. Start with the two rivals that come up most in Discord and sales conversations.

Each page catches a high-intent query and gives assistants a citable source. The format already exists, so each one is days of work.

• Rival facts go stale

  Date each page and link their docs.

• Your websiteSitemap (245 URLs, checked October 22, 2026): 165 changelog entries, 29 blog posts, 13 customer stories, 13 launch-week pages, 6 product pages; comparison pages only for Temporal, Inngest, BullMQ and n8n, plus "Claude Managed Agents vs Trigger.dev" and an agent-frameworks comparison. The homepage tabs list use cases (AI agents, Media processing, Media generation, Human in the loop, Streaming, Run Python, Marketing, Browser automation, Semantic search, Email sequences) but none has its own page. There is a /for-ai-crawlers page and an llms.txt, and every page is available as Markdown.Source
• Your websitellms.txt offers an "Agent setup skill" (trigger.dev/SKILL.md): "Instructions a coding agent (Claude Code, Cursor, Codex, and others) follows to sign in, create or choose a project, add Trigger.dev to a JavaScript or TypeScript app, and run a first task locally". The homepage has a "Copy setup prompt / Paste in your agent" button; an MCP server and "Ask Trigger" dashboard agent (beta, October 5) also shipped.Source
• Competitor researchRivals' own sites: Inngest, "Durable Execution for Workflows & AI"; Hatchet, "The orchestration engine for teams who ship", with "Hatchet vs Celery" and "Hatchet vs Temporal" pages; Restate and DBOS each publish a page against Temporal. None of these has a Trigger.dev comparison in its sitemap, so Vercel's page is the only rival comparison an assistant finds there.See the research Source
• Competitor researchVercel's Workflow SDK docs publish "Workflow SDK vs trigger.dev" (with "Copy for LLM" and "View Markdown" buttons). Its table lists trigger.dev limits as "3 MB payload / 10 MB output; 14-day maximum run lifetime", says "the 14-day maximum run time means human-in-the-loop flows that wait longer than two weeks can't complete in one run", that large model contexts "need external storage", and that trigger.dev is "TypeScript / JavaScript only". The same docs compare the Workflow SDK with Temporal, Cloudflare Workflows, AWS Step Functions, AWS Bedrock AgentCore and Inngest.See the research Source

Add one matching customer quote and number to each comparison page: the Magic Patterns migration on /vs/bullmq, an AI customer on /vs/inngest and /vs/temporal. Link the full story.

Proof at the moment of comparison lifts conversion more than proof on a page few comparers visit.

• Customer approval needed

  Use quotes already published in their stories.

• Your websiteCustomer story: "How Magic Patterns migrated 200k monthly jobs to Trigger.dev in one day", moving from "their unreliable self-hosted solution". Other stories include MagicSchool AI, Icon, Midday, Cal.com and Papermark.Source
• Your websiteSitemap (245 URLs, checked October 22, 2026): 165 changelog entries, 29 blog posts, 13 customer stories, 13 launch-week pages, 6 product pages; comparison pages only for Temporal, Inngest, BullMQ and n8n, plus "Claude Managed Agents vs Trigger.dev" and an agent-frameworks comparison. The homepage tabs list use cases (AI agents, Media processing, Media generation, Human in the loop, Streaming, Run Python, Marketing, Browser automation, Semantic search, Email sequences) but none has its own page. There is a /for-ai-crawlers page and an llms.txt, and every page is available as Markdown.Source

Add one contextual line under the viewer ("Debugging JSON from an AI agent? Trigger.dev traces every run and its payloads") with a tracked link to a landing page, an "Open in JSON Hero" button on run payloads in the dashboard, and a small paid search budget for "json viewer" queries pointing to JSON Hero. Measure signups with the tracked link.

A useful free tool earns attention without asking for it; one relevant line converts some of that attention into trials, and paid spend on a free tool is cheaper than on a signup page.

• Users dislike ads in a tool

  One quiet, relevant line, never a pop-up.

• GitHubtriggerdotdev/trigger.dev has 16.6k stars and about 1,500 forks (Apache 2.0); the careers page cites "5.1k+ Discord members". JSON Hero (jsonhero-web, 10.9k stars, last pushed November 28, 2025) is "Built with love by the Trigger.dev team"; jsonhero.io credits Trigger.dev only with a small "by" logo link and says nothing about what Trigger.dev does.Source
• HiringMarketing Lead (posted September 16, 2026, remote CET to ET or London): "own growth marketing at Trigger.dev: influencer partnerships, GEO/SEO, and the bold, high-visibility campaigns that get us noticed"; "picks one or two and goes deep"; influencer work "using tools like Plug.dev"; "big swing" campaigns such as "an SF billboard, or premium merch sends to YC founders"; and "narrow, targeted paid spend (e.g. a small daily budget tied to a free tool)".Source
• Your websitePricing: Free $0 with "$5 / month free credits", 20 concurrent runs and 1 day log retention; Hobby $10; Pro $50 with 30 day retention and "Dedicated Slack support"; Enterprise custom. Compute is billed per second (Small 1x $0.0000338/s) plus $0.25 per 10,000 runs, and the FAQ explains usage with a worked example ("1 month = $0.0363 x 30 = $1.09").Source

Write 30 prompts developers really ask ("run a long AI agent in TypeScript", "background jobs for Next.js", "BullMQ alternative"). Run them weekly in four assistants, log mentions, rank and wrong facts, and link each wrong fact to the page that should fix it.

You cannot improve answer share without measuring it, and wrong facts are cheaper to fix than missing mentions.

• Answers vary run to run

  Run each prompt three times and use the majority.

• HiringMarketing Lead (posted September 16, 2026, remote CET to ET or London): "own growth marketing at Trigger.dev: influencer partnerships, GEO/SEO, and the bold, high-visibility campaigns that get us noticed"; "picks one or two and goes deep"; influencer work "using tools like Plug.dev"; "big swing" campaigns such as "an SF billboard, or premium merch sends to YC founders"; and "narrow, targeted paid spend (e.g. a small daily budget tied to a free tool)".Source
• Your websiteSitemap (245 URLs, checked October 22, 2026): 165 changelog entries, 29 blog posts, 13 customer stories, 13 launch-week pages, 6 product pages; comparison pages only for Temporal, Inngest, BullMQ and n8n, plus "Claude Managed Agents vs Trigger.dev" and an agent-frameworks comparison. The homepage tabs list use cases (AI agents, Media processing, Media generation, Human in the loop, Streaming, Run Python, Marketing, Browser automation, Semantic search, Email sequences) but none has its own page. There is a /for-ai-crawlers page and an llms.txt, and every page is available as Markdown.Source
• Competitor researchVercel's Workflow SDK docs publish "Workflow SDK vs trigger.dev" (with "Copy for LLM" and "View Markdown" buttons). Its table lists trigger.dev limits as "3 MB payload / 10 MB output; 14-day maximum run lifetime", says "the 14-day maximum run time means human-in-the-loop flows that wait longer than two weeks can't complete in one run", that large model contexts "need external storage", and that trigger.dev is "TypeScript / JavaScript only". The same docs compare the Workflow SDK with Temporal, Cloudflare Workflows, AWS Step Functions, AWS Bedrock AgentCore and Inngest.See the research Source

Add a calculator: task length, runs per day, machine size and plan, showing monthly compute, run cost and which plan fits. Offer presets for common jobs (AI agent turn, video transcode, nightly cron).

A concrete monthly number removes the fear of a surprise bill, the most common worry with usage pricing.

• Underestimates cost

  Show a range and the free credits clearly.

• Your websitePricing: Free $0 with "$5 / month free credits", 20 concurrent runs and 1 day log retention; Hobby $10; Pro $50 with 30 day retention and "Dedicated Slack support"; Enterprise custom. Compute is billed per second (Small 1x $0.0000338/s) plus $0.25 per 10,000 runs, and the FAQ explains usage with a worked example ("1 month = $0.0363 x 30 = $1.09").Source
• HiringDesign Engineer (posted September 16, 2026): "raise the bar on UX and visual quality across our product and marketing surfaces", build "polished UI across our web app and marketing site", evolve the design system and component library, and improve "accessibility, responsiveness, and interaction polish". Stack: React, TypeScript, Remix.Source

Build ten use-case pages, each with the problem, a working example task, a customer quote where one exists, limits that matter and a "deploy this" button. Start with media processing, browser automation and Run Python.

Use-case searches are where developers start before they know which tool they need.

• Thin pages

  Every page needs a runnable example.

• Your websiteSitemap (245 URLs, checked October 22, 2026): 165 changelog entries, 29 blog posts, 13 customer stories, 13 launch-week pages, 6 product pages; comparison pages only for Temporal, Inngest, BullMQ and n8n, plus "Claude Managed Agents vs Trigger.dev" and an agent-frameworks comparison. The homepage tabs list use cases (AI agents, Media processing, Media generation, Human in the loop, Streaming, Run Python, Marketing, Browser automation, Semantic search, Email sequences) but none has its own page. There is a /for-ai-crawlers page and an llms.txt, and every page is available as Markdown.Source
• Your websiteCustomer story: "How Magic Patterns migrated 200k monthly jobs to Trigger.dev in one day", moving from "their unreliable self-hosted solution". Other stories include MagicSchool AI, Icon, Midday, Cal.com and Papermark.Source
• Your docsThe Python build extension: "If you need to execute Python scripts in your Trigger.dev project, you can use the pythonExtension build extension via the @trigger.dev/python package." The homepage use-case tabs include "Run Python".Source

Publish a "Python on Trigger.dev" page: orchestrate in TypeScript, run your Python scripts and models in the same task, with two examples (a transcription model, a scraping script). Be clear about what is TypeScript and what is Python.

It turns a stated weakness into a specific, honest strength for mixed-language teams.

• Overclaims Python support

  State plainly that tasks are written in TypeScript.

• Your docsThe Python build extension: "If you need to execute Python scripts in your Trigger.dev project, you can use the pythonExtension build extension via the @trigger.dev/python package." The homepage use-case tabs include "Run Python".Source
• Competitor researchVercel's Workflow SDK docs publish "Workflow SDK vs trigger.dev" (with "Copy for LLM" and "View Markdown" buttons). Its table lists trigger.dev limits as "3 MB payload / 10 MB output; 14-day maximum run lifetime", says "the 14-day maximum run time means human-in-the-loop flows that wait longer than two weeks can't complete in one run", that large model contexts "need external storage", and that trigger.dev is "TypeScript / JavaScript only". The same docs compare the Workflow SDK with Temporal, Cloudflare Workflows, AWS Step Functions, AWS Bedrock AgentCore and Inngest.See the research Source
• Your websiteSitemap (245 URLs, checked October 22, 2026): 165 changelog entries, 29 blog posts, 13 customer stories, 13 launch-week pages, 6 product pages; comparison pages only for Temporal, Inngest, BullMQ and n8n, plus "Claude Managed Agents vs Trigger.dev" and an agent-frameworks comparison. The homepage tabs list use cases (AI agents, Media processing, Media generation, Human in the loop, Streaming, Run Python, Marketing, Browser automation, Semantic search, Email sequences) but none has its own page. There is a /for-ai-crawlers page and an llms.txt, and every page is available as Markdown.Source

Give each creator a starter repo (a durable chat agent or a video pipeline), a tracked signup link with extra credits, and one brief: build it on camera, show a crash and resume. Pay on a flat fee plus signups, and reuse the best clips.

One format makes results comparable between creators and shows the product's best trick, surviving crashes and redeploys.

• Vanity reach, no activation

  Pay partly on deployed tasks.

• HiringMarketing Lead (posted September 16, 2026, remote CET to ET or London): "own growth marketing at Trigger.dev: influencer partnerships, GEO/SEO, and the bold, high-visibility campaigns that get us noticed"; "picks one or two and goes deep"; influencer work "using tools like Plug.dev"; "big swing" campaigns such as "an SF billboard, or premium merch sends to YC founders"; and "narrow, targeted paid spend (e.g. a small daily budget tied to a free tool)".Source
• Your blogSeries A (December 17, 2025): "$16M Series A led by Standard Capital", used "to execute hundreds of millions of agents each month by over 30,000 developers". Next: "more advanced observability", "faster run starts using MicroVMs", "more integrations with common third-party services", "Sandboxes will allow you to execute untrusted code", and "a whole suite of tools for context engineering".Source
• Your websitellms.txt offers an "Agent setup skill" (trigger.dev/SKILL.md): "Instructions a coding agent (Claude Code, Cursor, Codex, and others) follows to sign in, create or choose a project, add Trigger.dev to a JavaScript or TypeScript app, and run a first task locally". The homepage has a "Copy setup prompt / Paste in your agent" button; an MCP server and "Ask Trigger" dashboard agent (beta, October 5) also shipped.Source

Each talk gets a public repo with the demo, a short README and a tracked "deploy in one command" link, plus a QR on the last slide. Log stars, forks and signups per talk next to the events report.

It keeps talks useful and non-salesy while showing which rooms create users.

• Feels like a pitch

  The repo is useful without Trigger.dev mentioned in the talk.

• HiringDevRel (SF Events), posted September 16, 2026: "100% events, not product education, content strategy, or general marketing"; show up at "meetups, AI Tinkers, YC events, conferences", win speaking slots, and over time organize events such as "joint hackathons at the YC offices". "The goal is pure awareness"; reporting is "number of events attended, video of any talks given, and a written recap of who you met".Source
• Your websitellms.txt offers an "Agent setup skill" (trigger.dev/SKILL.md): "Instructions a coding agent (Claude Code, Cursor, Codex, and others) follows to sign in, create or choose a project, add Trigger.dev to a JavaScript or TypeScript app, and run a first task locally". The homepage has a "Copy setup prompt / Paste in your agent" button; an MCP server and "Ask Trigger" dashboard agent (beta, October 5) also shipped.Source

Co-host with one AI model provider and one framework partner; theme it "agents that keep running"; give every team Pro credits; judge on reliability under a forced crash. Publish the winning repos.

Hackathons create hands-on users and content in one day, and the theme shows your difference.

• Low turnout

  Co-hosts promote to their lists.

• HiringDevRel (SF Events), posted September 16, 2026: "100% events, not product education, content strategy, or general marketing"; show up at "meetups, AI Tinkers, YC events, conferences", win speaking slots, and over time organize events such as "joint hackathons at the YC offices". "The goal is pure awareness"; reporting is "number of events attended, video of any talks given, and a written recap of who you met".Source
• Your blogSeries A (December 17, 2025): "$16M Series A led by Standard Capital", used "to execute hundreds of millions of agents each month by over 30,000 developers". Next: "more advanced observability", "faster run starts using MicroVMs", "more integrations with common third-party services", "Sandboxes will allow you to execute untrusted code", and "a whole suite of tools for context engineering".Source

Send founders of recent YC companies with background-job needs a quality item plus a card with a personal link: a free migration session with an engineer and extra credits. Track bookings by batch.

Pairing a gift with a specific, low-effort offer turns goodwill into meetings.

• Feels spammy

  Only companies that fit, with a personal note.

• HiringMarketing Lead (posted September 16, 2026, remote CET to ET or London): "own growth marketing at Trigger.dev: influencer partnerships, GEO/SEO, and the bold, high-visibility campaigns that get us noticed"; "picks one or two and goes deep"; influencer work "using tools like Plug.dev"; "big swing" campaigns such as "an SF billboard, or premium merch sends to YC founders"; and "narrow, targeted paid spend (e.g. a small daily budget tied to a free tool)".Source
• Your websiteCustomer story: "How Magic Patterns migrated 200k monthly jobs to Trigger.dev in one day", moving from "their unreliable self-hosted solution". Other stories include MagicSchool AI, Icon, Midday, Cal.com and Papermark.Source
• Your blogSeries A (December 17, 2025): "$16M Series A led by Standard Capital", used "to execute hundreds of millions of agents each month by over 30,000 developers". Next: "more advanced observability", "faster run starts using MicroVMs", "more integrations with common third-party services", "Sandboxes will allow you to execute untrusted code", and "a whole suite of tools for context engineering".Source

Measure signups and first runs that come through SKILL.md and the MCP server, find where agents fail (sign-in, config, first deploy), and fix the top three. Publish "add Trigger.dev with your agent" guides for each coding tool.

If the agent completes setup, activation stops depending on the developer reading docs.

• Agent tools change fast

  Retest monthly.

• Your websitellms.txt offers an "Agent setup skill" (trigger.dev/SKILL.md): "Instructions a coding agent (Claude Code, Cursor, Codex, and others) follows to sign in, create or choose a project, add Trigger.dev to a JavaScript or TypeScript app, and run a first task locally". The homepage has a "Copy setup prompt / Paste in your agent" button; an MCP server and "Ask Trigger" dashboard agent (beta, October 5) also shipped.Source
• HiringEvery Trigger.dev job post: "We are a product-led growth company so everyone is expected to get involved in creating content like code examples, blog articles, videos and tweets" and "Great documentation matters more than ever, both for humans and for the AI tools that read it." The Senior Backend Engineer post says the Cloud serves "thousands of teams building AI apps and agents, handling hundreds of millions of executions per month"; a Backend Engineer (Security) role is also open.Source

When a user opens an expired run or hits the concurrency cap, show a clear in-dashboard message with what Hobby or Pro would give, and a one-click upgrade. Track upgrades per trigger.

Prompts tied to a real need convert better than generic upgrade emails.

• Annoys users

  Show once per day at most.

• Your websitePricing: Free $0 with "$5 / month free credits", 20 concurrent runs and 1 day log retention; Hobby $10; Pro $50 with 30 day retention and "Dedicated Slack support"; Enterprise custom. Compute is billed per second (Small 1x $0.0000338/s) plus $0.25 per 10,000 runs, and the FAQ explains usage with a worked example ("1 month = $0.0363 x 30 = $1.09").Source

Build /reliability: uptime by component, every incident report with its fix, architecture notes (checkpointing, versioning) and security and compliance links (SOC 2, HIPAA).

Buyers evaluating infrastructure look for how a vendor handles failure, and openness wins trust.

• Highlights failures

  Pair each with the fix and the trend.

• Your blogRecent posts are engineering-led: "7 things we wish we'd known before building an AI chat agent" (October 14), "Zod 4 got rid of a bunch of CPUs for us", "Upgrading a 56TB database that's growing at 250GB per day", incident reports for June 30 and October 6, 2026, and "10 Claude Code tips you did not know".Source
• HiringEvery Trigger.dev job post: "We are a product-led growth company so everyone is expected to get involved in creating content like code examples, blog articles, videos and tweets" and "Great documentation matters more than ever, both for humans and for the AI tools that read it." The Senior Backend Engineer post says the Cloud serves "thousands of teams building AI apps and agents, handling hundreds of millions of executions per month"; a Backend Engineer (Security) role is also open.Source

Publish "move from self-hosted to Cloud in an afternoon" with export steps, offer credits for the first month, and mention it in self-hosting docs and release notes.

Self-hosters already know the product; moving them is the cheapest new revenue.

• Upsets OSS community

  Keep self-hosting first-class and the tone helpful.

• GitHubtriggerdotdev/trigger.dev has 16.6k stars and about 1,500 forks (Apache 2.0); the careers page cites "5.1k+ Discord members". JSON Hero (jsonhero-web, 10.9k stars, last pushed November 28, 2025) is "Built with love by the Trigger.dev team"; jsonhero.io credits Trigger.dev only with a small "by" logo link and says nothing about what Trigger.dev does.Source
• Your websiteCustomer story: "How Magic Patterns migrated 200k monthly jobs to Trigger.dev in one day", moving from "their unreliable self-hosted solution". Other stories include MagicSchool AI, Icon, Midday, Cal.com and Papermark.Source
• Your websiteSitemap (245 URLs, checked October 22, 2026): 165 changelog entries, 29 blog posts, 13 customer stories, 13 launch-week pages, 6 product pages; comparison pages only for Temporal, Inngest, BullMQ and n8n, plus "Claude Managed Agents vs Trigger.dev" and an agent-frameworks comparison. The homepage tabs list use cases (AI agents, Media processing, Media generation, Human in the loop, Streaming, Run Python, Marketing, Browser automation, Semantic search, Email sequences) but none has its own page. There is a /for-ai-crawlers page and an llms.txt, and every page is available as Markdown.Source

Run a billboard near developer offices with a short line about agents that survive crashes and a vanity URL that opens a challenge (fix a task that fails, deploy it). Solvers get credits and a shirt. Time it with a launch week.

A puzzle turns a broadcast ad into a hands-on trial and gives people a reason to share it.

• Low response

  Pair with social and events in the same week.

• HiringMarketing Lead (posted September 16, 2026, remote CET to ET or London): "own growth marketing at Trigger.dev: influencer partnerships, GEO/SEO, and the bold, high-visibility campaigns that get us noticed"; "picks one or two and goes deep"; influencer work "using tools like Plug.dev"; "big swing" campaigns such as "an SF billboard, or premium merch sends to YC founders"; and "narrow, targeted paid spend (e.g. a small daily budget tied to a free tool)".Source
• HiringDevRel (SF Events), posted September 16, 2026: "100% events, not product education, content strategy, or general marketing"; show up at "meetups, AI Tinkers, YC events, conferences", win speaking slots, and over time organize events such as "joint hackathons at the YC offices". "The goal is pure awareness"; reporting is "number of events attended, video of any talks given, and a written recap of who you met".Source
• Your blogSeries A (December 17, 2025): "$16M Series A led by Standard Capital", used "to execute hundreds of millions of agents each month by over 30,000 developers". Next: "more advanced observability", "faster run starts using MicroVMs", "more integrations with common third-party services", "Sandboxes will allow you to execute untrusted code", and "a whole suite of tools for context engineering".Source

Publish a reproducible benchmark (cold start, checkpoint and resume time) with code anyone can run, launch Sandboxes with its own page and launch week, and update every comparison page with the numbers.

Hard numbers travel further than features and give assistants something concrete to cite.

• Benchmark disputes

  Publish code and invite reruns.

• Your blogSeries A (December 17, 2025): "$16M Series A led by Standard Capital", used "to execute hundreds of millions of agents each month by over 30,000 developers". Next: "more advanced observability", "faster run starts using MicroVMs", "more integrations with common third-party services", "Sandboxes will allow you to execute untrusted code", and "a whole suite of tools for context engineering".Source
• Your websiteSitemap (245 URLs, checked October 22, 2026): 165 changelog entries, 29 blog posts, 13 customer stories, 13 launch-week pages, 6 product pages; comparison pages only for Temporal, Inngest, BullMQ and n8n, plus "Claude Managed Agents vs Trigger.dev" and an agent-frameworks comparison. The homepage tabs list use cases (AI agents, Media processing, Media generation, Human in the loop, Streaming, Run Python, Marketing, Browser automation, Semantic search, Email sequences) but none has its own page. There is a /for-ai-crawlers page and an llms.txt, and every page is available as Markdown.Source
• Competitor researchRivals' own sites: Inngest, "Durable Execution for Workflows & AI"; Hatchet, "The orchestration engine for teams who ship", with "Hatchet vs Celery" and "Hatchet vs Temporal" pages; Restate and DBOS each publish a page against Temporal. None of these has a Trigger.dev comparison in its sitemap, so Vercel's page is the only rival comparison an assistant finds there.See the research Source

Publish an annual report from aggregated data: run lengths, waits, retries, models used, human-in-the-loop share. Launch it with talks and a blog series.

Original data earns links, press and assistant citations, and positions Trigger.dev as the authority on production agents.

• Customer privacy

  No customer identifiable; suppress small groups.

• Your blogSeries A (December 17, 2025): "$16M Series A led by Standard Capital", used "to execute hundreds of millions of agents each month by over 30,000 developers". Next: "more advanced observability", "faster run starts using MicroVMs", "more integrations with common third-party services", "Sandboxes will allow you to execute untrusted code", and "a whole suite of tools for context engineering".Source
• HiringEvery Trigger.dev job post: "We are a product-led growth company so everyone is expected to get involved in creating content like code examples, blog articles, videos and tweets" and "Great documentation matters more than ever, both for humans and for the AI tools that read it." The Senior Backend Engineer post says the Cloud serves "thousands of teams building AI apps and agents, handling hundreds of millions of executions per month"; a Backend Engineer (Security) role is also open.Source
• Your blogRecent posts are engineering-led: "7 things we wish we'd known before building an AI chat agent" (October 14), "Zod 4 got rid of a bunch of CPUs for us", "Upgrading a 56TB database that's growing at 250GB per day", incident reports for June 30 and October 6, 2026, and "10 Claude Code tips you did not know".Source

Release a small React kit (run progress, streaming output, approval button for human-in-the-loop) built on your design system, with copy-paste install and a demo page.

Free components spread through code, and each install is a Realtime adoption.

• Maintenance load

  Keep it small.

• HiringDesign Engineer (posted September 16, 2026): "raise the bar on UX and visual quality across our product and marketing surfaces", build "polished UI across our web app and marketing site", evolve the design system and component library, and improve "accessibility, responsiveness, and interaction polish". Stack: React, TypeScript, Remix.Source
• Your websiteSitemap (245 URLs, checked October 22, 2026): 165 changelog entries, 29 blog posts, 13 customer stories, 13 launch-week pages, 6 product pages; comparison pages only for Temporal, Inngest, BullMQ and n8n, plus "Claude Managed Agents vs Trigger.dev" and an agent-frameworks comparison. The homepage tabs list use cases (AI agents, Media processing, Media generation, Human in the loop, Streaming, Run Python, Marketing, Browser automation, Semantic search, Email sequences) but none has its own page. There is a /for-ai-crawlers page and an llms.txt, and every page is available as Markdown.Source

Keep one engineering post a month and add one practical "how to" a month tied to a use-case page (for example "long-running AI agents on Vercel without timeouts"). Track which posts drive signups.

Matching content to buyer searches makes the blog a growth channel, not just a reputation channel.

• Dilutes the engineering voice

  Keep engineers as authors.

• Your blogRecent posts are engineering-led: "7 things we wish we'd known before building an AI chat agent" (October 14), "Zod 4 got rid of a bunch of CPUs for us", "Upgrading a 56TB database that's growing at 250GB per day", incident reports for June 30 and October 6, 2026, and "10 Claude Code tips you did not know".Source
• HiringEvery Trigger.dev job post: "We are a product-led growth company so everyone is expected to get involved in creating content like code examples, blog articles, videos and tweets" and "Great documentation matters more than ever, both for humans and for the AI tools that read it." The Senior Backend Engineer post says the Cloud serves "thousands of teams building AI apps and agents, handling hundreds of millions of executions per month"; a Backend Engineer (Security) role is also open.Source
• Your websiteSitemap (245 URLs, checked October 22, 2026): 165 changelog entries, 29 blog posts, 13 customer stories, 13 launch-week pages, 6 product pages; comparison pages only for Temporal, Inngest, BullMQ and n8n, plus "Claude Managed Agents vs Trigger.dev" and an agent-frameworks comparison. The homepage tabs list use cases (AI agents, Media processing, Media generation, Human in the loop, Streaming, Run Python, Marketing, Browser automation, Semantic search, Email sequences) but none has its own page. There is a /for-ai-crawlers page and an llms.txt, and every page is available as Markdown.Source

Set a performance budget for marketing pages (LCP under 2 s at the 75th percentile), check it in CI with Lighthouse, and review heavy assets before release.

Developers notice slow sites, and a budget protects the speed you already have.

• Blocks releases

  Warn first, block later.

• Chrome UX ReportReal visitors to trigger.dev pass Core Web Vitals: largest contentful paint 1,537 ms and interaction to next paint 74 ms at the 75th percentile, though LCP has slowed about 11% over the period.Source
• HiringDesign Engineer (posted September 16, 2026): "raise the bar on UX and visual quality across our product and marketing surfaces", build "polished UI across our web app and marketing site", evolve the design system and component library, and improve "accessibility, responsiveness, and interaction polish". Stack: React, TypeScript, Remix.Source