FAQ: Getting Started & Common Setup Traps

[satur8d]

Getting set up with Totem? Here are a few common questions and traps we've seen during the early alpha:

Q: Do I run totem init in the Totem repo itself?
A: No! You run npx @mmnto/cli init inside the consuming project (e.g., your Next.js app or Node backend). Totem is designed to live inside the project it is analyzing.

Q: I'm hitting strict OpenAI rate limits during my first totem sync.
A: If you have a massive codebase, the initial full index can hit Tier 1 OpenAI limits. We recommend changing your provider in totem.config.ts to use local Ollama (nomic-embed-text) for the initial sync to avoid costs/limits, then switch back to OpenAI for faster daily use.

Q: I manually deleted my .lancedb folder and now sync is broken.
A: If you delete the folder, incremental sync gets confused. Run pnpm exec totem sync --full to force a complete recreation of the index from scratch.

Q: Does Totem upload my code to the cloud?
A: No. Your memory index (.lancedb/) lives entirely locally. If you use OpenAI for embeddings, chunks are sent there briefly for processing, but you can configure Totem to use 100% local Ollama for total privacy.

Have a question that isn't answered here? Start a new thread in the Q&A category!