recall

Index past Claude Code and Codex conversations into a SQLite FTS5 database.

284|10|Updated Dec 6, 2015

Search Tags:#codex #search #claude #retrieval #history #conversations

Ready to run

npx skills add https://github.com/meain/dotfiles --skill recall-meain

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: recall
Source: https://github.com/meain/dotfiles/tree/main/agents/.agents/skills/recall
Command: npx skills add https://github.com/meain/dotfiles --skill recall-meain

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill helps you quickly find information buried within past conversation sessions with AI assistants like Claude Code and Codex, preventing you from losing valuable context or having to repeat yourself.

Core Features & Use Cases

Full-Text Search: Indexes and searches across all your past conversation transcripts.
Cross-Platform: Works with both Claude Code and Codex session data.
Intelligent Ranking: Uses BM25 ranking with a recency boost to surface the most relevant results.
Use Case: You remember discussing a specific API integration a few weeks ago but can't recall the exact details. Use this Skill to search for keywords related to the API and instantly retrieve the relevant conversation snippet and session file.

Quick Start

Search for past conversations about 'bufferStore' using the recall skill.

Frequently Asked Questions about recall

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I search past Claude Code and Codex conversations for specific code snippets?▼

You can search past Claude Code and Codex conversations by indexing user and assistant messages into a SQLite FTS5 database. This enables full-text retrieval of specific code snippets, discussion points, or decisions from prior sessions.

Can I retrieve old AI assistant chat history using full-text search?▼

Yes, you can retrieve old AI assistant chat history using full-text search. The skill indexes past conversation transcripts from Claude Code and Codex, allowing you to query and surface relevant past interactions instantly.

How does BM25 ranking work when searching past conversation sessions?▼

BM25 ranking in past conversation search surfaces the most relevant results by matching keyword frequency. This mechanism applies a recency boost, ensuring that newer discussions about your query terms rank higher in the retrieved results.

Does this conversation retrieval tool work with both Claude Code and Codex session data?▼

Yes, this conversation retrieval tool works cross-platform with both Claude Code and Codex session data. It indexes transcripts from both platforms into a unified SQLite FTS5 database for seamless full-text search.

What is the best way to find a specific API integration discussed in previous AI chat sessions?▼

The best way to find a specific API integration discussed in previous AI chat sessions is to search for related keywords using full-text retrieval. The system matches your query against indexed past messages and returns the relevant conversation snippet and session file.

Do I need to manually update the search index after new Claude Code or Codex conversations?▼

No, you do not need to manually update the search index after new conversations. The system supports incremental updates, automatically indexing new user and assistant messages into the SQLite FTS5 database as your past conversation history grows.