Find AI builders by proof, not talk.
Search by what they actually build — tokens, builds, hours, tools — captured live from real coding sessions.
Building with AI yourself? Get found →
229 builders · 19 projects · 6 open to opportunities
229 builders, 19 projects, 6 open to opportunities.One-shot session on tmp
cheating-daddy
LOCKED IN
Back after 8 days
Welcome back
737.7ktokens
10prompts
5:51:59time
1,281lines
Claude Codeclaude-sonnet-5Markdown$11.01
mostly building
13s coding session using gpt-6-astra. 62K tokens across 1 prompts.
One-shot session on tmp
eltempsavui
Longest session this week
12h 24m
462.9ktokens
3prompts
12:24:47time
8lines
Claude Codeclaude-opus-5TypeScript$12.56
mostly verifying
13s coding session using gpt-6-astra. 62K tokens across 1 prompts.
Audited site voice and design system foundations
claude-partner-network
DEEP WORK
Top 5% session this week
802 sessions tracked this week across all builders
3.8Mtokens
12prompts
7:53:11time
2,795lines
Claude Codeclaude-fable-5-1$85.18
mostly building
Expert panel reviewed website language; design audit completed with 6 verification lenses and system foundations documented for build.
Shipped dark and light modes for documentation site
flownest-dev
MARATHON
Most lines changed this week
Previous best: 308 lines
1.5Mtokens
34prompts
4:51:29time
674lines
Claude Codeclaude-opus-5[1m]TypeScript$106.67
Deployed dark and light theme modes for the documentation site, verified theme persistence on staging.
Founding Engineer @ Matchory. Passionate OSS dev with 15y experience. Sucker for type safety, elegant abstractions, and DevEx.
5886.0Mtokens
2.1k
builds
19638.8h
build time
31d
active 30d
Verified SBOM sync automation across all execution paths
Cetacean
LOCKED IN
Top 17% session this week
800 sessions tracked this week across all builders
708.9ktokens
4prompts
3:03:39time
80lines
Claude Codeclaude-opus-5Shell$36.99
mostly verifying
Finished automation for SBOM sync, tested both execution paths, and verified all linting checks.
Planned the Gateway vision strategy with site audit and council feedback
claude-partner-network
MARATHON
Most lines changed this week
Previous best: 1,112 lines
10.0Mtokens
62prompts
44:17:55time
3,062lines
Claude Codeclaude-fable-5-1Markdown$227.77
balanced mix
Audited the Gateway site, gathered council design feedback, and documented the vision strategy plan.
Dirigible
Voxel based colony sim/god game.
2.3Btokens
638
builds
7.4k
prompts
1259.9h
build time
One-shot debugging on write
rewire
Top 17% session this week
794 sessions tracked this week across all builders
675.2ktokens
10prompts
1:19:40time
784lines
Claude Codeclaude-fable-5Swift$61.75
20h 24m session running commands and testing. 87K tokens across 1 prompts using claude-opus-5.
Verified and pushed seven bug fixes to PR
Cetacean
FULL SEND
5.3x your typical session
2.2M tokens vs typical 416.0K
2.2Mtokens
5prompts
7:24:40time
168lines
Claude Codeclaude-opus-5[1m]Markdown$79.56
mostly verifying
Verified seven bug fixes in cluster operations and pushed all commits to complete the PR.
Built working Docker Compass setup with local and web variants
mongo-compass-docker
ONE-SHOT
First session with fable-5-1
First build using this model
372.1ktokens
1prompts
34:49time
Claude Codeclaude-fable-5-1$10.48
mostly verifying
Rebuilt and tested Docker Compose files for local desktop and web-based Compass deployments.
Résumés tell. Promptbook shows.
Every other platform asks builders to describe themselves. We capture what they actually did.
The old way
With Promptbook
Self-written portfolio
Captured build activity
Polished résumé claims
Live sessions from real tools
Static project list
Recent build velocity
Manual screenshots
Server-rendered build cards
Guess who's active
Search by recency, tools, tokens
How it works
01
Builders connect their tools
One install hooks Claude Code or Codex. Every coding session is captured automatically — no uploads, no self-reporting.
02
Sessions become proof
Tokens, builds, hours, tools and languages turn into a searchable profile and shareable build cards. Never their code or prompts.
03
Companies search the real ones
Filter by what builders actually ship — then reach out. Browsing is free; you pay only to contact.
For companies & investors
Find builders by proof, not promises. Search the full pool free — pay only when you want to reach out.
Browse builders →For builders
Get found for the AI work you're already doing. Free, forever — builders never pay for visibility, rank, or placement.
Get found →Captured live
Streamed from real coding sessions. Nothing self-reported, nothing to upload.
Private by default
Never your source code or prompts. You control which projects are visible.
Tokens, not talk
The honest measure of real AI building — the hardest stat to game.
The database of verified AI builders.
229 builders and counting, 6 open to opportunities.