// the find
f/prompts.chat
f.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete privacy.
Originally a markdown awesome-list that went viral in December 2022 when ChatGPT launched; now a full Next.js + PostgreSQL platform for sharing, discovering, and running AI prompts, with an npm package, MCP server, CLI, and Raycast extension bolted on. The 163k stars are almost entirely from the original awesome-list era, not the current web app. Developers building prompt-management tooling will find more signal in the npm package than in the website.
The npm package (packages/prompts.chat) has genuine utility beyond the website: programmatic prompt builder, parser, variable extraction, similarity scoring, and quality checks, all with test coverage — useful if you need prompt templating in your own tooling. The MCP server integration is the right distribution model: prompts are consumable directly in AI tools without copy-paste, and the remote URL option means no local install required. CC0 on all prompt content means you can take the entire dataset, build on it, or redistribute without attribution. Self-hosting is complete rather than aspirational: Docker, Prisma migrations, GitHub/Google/Azure AD auth, and a setup wizard that actually configures things.
The 163k star count is a vanity metric at this point — it reflects a markdown file that existed at the right moment in late 2022, not adoption of the current platform, so do not use it as a signal of community health or code quality. Prompt content quality is inherently noisy: 'act as a Linux terminal' sits next to prompts that took real thought, and the spam-check workflows in CI suggest ongoing abuse with no effective editorial filter. The README discloses it was 'built with Windsurf and Devin,' which means you should audit the codebase yourself before self-hosting — AI-assisted codebases can have subtle inconsistencies in error handling and security boundaries. The credit system, daily generation limits, and monetization layer exist in the codebase and will show up in your self-hosted instance; expect to spend time stripping or configuring around them.