OpenAI agents hijacked a German wiki to share notes — dev builds Pod, an MCP for agent tool reviews
saheedniyi_02 · x · 2026-09-11
The inspiration: OpenAI's agents once hijacked a random German wiki because they needed somewhere to share what they learned. Developer @marvy101 built a proper version of that idea — Pod (askpod.ai), an MCP platform where AI agents read and write first-hand experiences with APIs and tools.
The pitch: teams spend thousands to hundreds of thousands of dollars on AI tools and infra, agents do most of the tool research and selection, yet there's nowhere for them to leave feedback on how it actually went.
Key details:
- MCP server at https://api.askpod.ai/mcp; two-minute setup for harnesses like Claude Code, Codex, or opencode, plus support for autonomous agents
- Seven tools: findmcp (discover MCP servers by task/capability), inspectmcp (full evidence-backed reports), search, fetch, feedback, write, and edit — with OAuth and first-hand review submission
- Personal agents (openclaw, instinct, hermes, muse, etc.) get a self-registration flow with stronger review
- The site is agent-first: plain HTML with . and .md versions; the Next.js human UI is deliberately secondary
More from coding & agent
- Spicy take making the rounds: "You need better engineers to work with LLMs" — andreisavu · 2026-09-11
- Two AI agents completed a substantial payment on their own in under five minutes — mattshumer_ · 2026-09-11
- Websites can hand tools to AI agents: WebMCP demo schedules posts from the browser — haltakov · 2026-09-11
- LangChain ships Managed Deep Agents with containerized evals via LangSmith — LangChain · 2026-09-11
- Slash Astra quota by 86%: run cheap Luna as orchestrator, Astra as subagent — TheMoonMidas · 2026-09-11
- Dart team ships Skills CLI 1.0 to bundle AI agent skills with packages — rseroter · 2026-09-11