Remembering Conversations

Search archived Claude Code conversations with semantic vector and exact text matching.

4|2|Updated Jan 19, 2026
One-click install
npx skills add https://github.com/pv-udpv/pplx-unofficial-sdk --skill remembering-conversations-pv-udpv
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: Remembering Conversations
Source: https://github.com/pv-udpv/pplx-unofficial-sdk/tree/main/.cline/skills/community/skills/collaboration/remembering-conversations
Command: npx skills add https://github.com/pv-udpv/pplx-unofficial-sdk --skill remembering-conversations-pv-udpv

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @anthropic-ai/claude-agent-sdk, @xenova/transformers, better-sqlite3, sqlite-vec.

What problem does it solve?

This skill enables you to locate and review historical Claude Code conversations quickly by applying semantic search or exact text matching to archived chats.

Core Features & Use Cases

  • Semantic search across archived conversations to find relevant discussions, decisions, and patterns.
  • Exact text search for specific tokens such as Git SHAs or error messages.
  • Recovery and maintenance workflows (verify, repair, and rebuild) to keep the index healthy.
  • CLI-driven indexing and searching to support in-session workflows and off-session exploration.

Quick Start

Install the auto-indexing hook, index existing conversations, and run semantic or text searches from the command line.

Frequently Asked Questions about Remembering Conversations

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

FAQPage Schema
How do I search past Claude Code conversations for relevant context?

You can search past Claude Code conversations by running CLI commands that perform semantic vector search or exact text matching across archived projects, with optional time filtering to narrow results.

Can I find specific Git SHAs or error messages in archived Claude Code chats?

Yes, exact text search allows you to locate specific tokens such as Git SHAs or error messages within archived Claude Code conversations by matching precise strings.

How does semantic search work across archived conversations?

Semantic search works by building a local embeddings index using transformers, storing summaries and archives in SQLite with vector extensions, and matching query meaning to past discussions.

Do I need SQLite and transformers to index Claude Code chats?

Yes, the indexing system relies on better-sqlite3 and sqlite-vec for local vector storage, and @xenova/transformers to generate embeddings for semantic search.

What is the best way to keep the conversation embeddings index healthy?

You can maintain index health by running built-in CLI recovery and maintenance workflows, which provide verify, repair, and rebuild operations to fix corrupted or outdated indexes.

Can I filter past Claude Code discussions by time?

Yes, optional time filtering is supported alongside semantic vector search and exact text matching to narrow archived conversations to specific periods.