Context:
The migration to gemini-embedding-001 drastically changed the initial setup requirements due to strict daily rate limits. We have implemented a "Zero-Setup" strategy where the Knowledge Base is downloaded automatically via npm install (using the prepare script).
Problem:
.github/AI_QUICK_START.md is outdated and misleading:
- It claims the "free tier is generous" (it's now very restrictive for full syncs).
- It states embedding "takes 2-5 minutes" (it now takes ~25 mins if done manually).
- It instructs users to rely on automatic embedding via MCP servers, which we want to avoid for the initial run.
Objective:
Rewrite .github/AI_QUICK_START.md to reflect the new workflow:
- Automatic Setup: Explain that
npm install downloads the pre-built Knowledge Base artifact.
- Manual Download: Document
npm run ai:download-kb for troubleshooting.
- API Key: Clarify that the Gemini API Key is primarily for the Agent interaction and incremental KB updates, not strictly required for the initial KB existence anymore.
- Rate Limits: Update the "Note on Cost" to mention the strict daily limits if users choose to rebuild the KB manually (
npm run ai:sync-kb).
Acceptance Criteria:
- The guide accurately guides a new user through the "Zero-Setup" process.
- Misleading claims about speed and rate limits are corrected.
Context: The migration to
gemini-embedding-001drastically changed the initial setup requirements due to strict daily rate limits. We have implemented a "Zero-Setup" strategy where the Knowledge Base is downloaded automatically vianpm install(using thepreparescript).Problem:
.github/AI_QUICK_START.mdis outdated and misleading:Objective: Rewrite
.github/AI_QUICK_START.mdto reflect the new workflow:npm installdownloads the pre-built Knowledge Base artifact.npm run ai:download-kbfor troubleshooting.npm run ai:sync-kb).Acceptance Criteria: