Recall FAQs

Question 1

What are the system requirements for running Recall?

Accepted Answer

Recall primarily runs on macOS due to its LaunchAgent setup for background indexing. It requires Node.js 18+, Ollama (with the `nomic-embed-text` model), and `uv` to manage ChromaDB. The Granola desktop app is only necessary if you wish to index meeting transcripts.

Question 2

Does Recall require an internet connection or any API keys?

Accepted Answer

No, Recall is built for complete offline functionality. It does not need an internet connection, external API keys, or any cloud-based services to perform its semantic search and indexing tasks.

Question 3

What types of documents can Recall index and search?

Accepted Answer

Recall is designed to index your local markdown files (recursively from a specified folder) and transcripts generated by the Granola desktop app. This allows you to search across your notes, documentation, and meeting discussions.

Question 4

How does Recall ensure privacy and data security?

Accepted Answer

Recall operates entirely offline. It uses local models (Ollama with `nomic-embed-text`) for text embeddings and stores all vector data in a local ChromaDB instance. No data ever leaves your machine, and it requires no API keys or cloud services, ensuring maximum privacy.

Question 5

What is Recall?

Accepted Answer

Recall is a local semantic search server that indexes your markdown documents and Granola meeting transcripts. It allows AI models, like Claude Code, to find relevant content by meaning, not just keywords, operating entirely on your machine without cloud services.

Recall

Recall

主要功能

使用案例

主要功能

使用案例