zzallirog/weighted-compact
by Various
Substrate-based context compaction for Claude Code. Inspectable signal weights, local-only, MCP-ready. Alternative to /compact + hook-based capture (claude-mem, etc.).
MCP
zzallirog/weighted-compact
Added 7 June 2026
Overview
Provides context compaction for Claude Code using a Substrate-based approach. It applies inspectable signal weights to prioritize information, runs entirely locally, and is compatible with the Model Context Protocol (MCP). Serves as an alternative to the built-in /compact command or hook-based methods like claude-mem.
Best for
Best for
Developers using Claude Code who want transparent, local context compaction
Use cases
- Compacting conversation context for Claude Code sessions
- Inspecting weight-assigned signals to understand compaction decisions
- Integrating local context compaction into MCP-compatible workflows
How to use
Install
pipx install 'weighted-compact[mcp]' Tools exposed
weighted-compact recapweighted-compact bootstrapweighted-compact mcp-serveweighted-compact rem-pass
Tested with
Claude Desktop, Claude Code
Notes
Provides context compaction for Claude Code using a Substrate-based approach. It applies inspectable signal weights to prioritize information, runs entirely locally, and is compatible with the Model Context Protocol (MCP). Serves as an alternative to the built-in /compact command or hook-based methods like claude-mem.
1 stars on GitHub. Last updated 2026-06-07. Licensed MIT.
Use cases
- Compacting conversation context for Claude Code sessions
- Inspecting weight-assigned signals to understand compaction decisions
- Integrating local context compaction into MCP-compatible workflows
Pros
- Runs locally with no external service dependencies
- Inspectable signal weights provide transparency into compression logic
- MCP-ready enables integration with other tools and interfaces
Cons
- Single GitHub star indicates minimal adoption and testing
- Requires familiarity with Substrate concepts for effective use
- Alternative to established tools, may lack community support and documentation
Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.
Pros
- Runs locally with no external service dependencies
- Inspectable signal weights provide transparency into compression logic
- MCP-ready enables integration with other tools and interfaces
Cons
- Single GitHub star indicates minimal adoption and testing
- Requires familiarity with Substrate concepts for effective use
- Alternative to established tools, may lack community support and documentation
Pairs with
Other entries in the index that connect to this one. Click through to see the chain.
Get the free Developer’s Field Guide
A 27-page field guide to the AI coding workflow with Claude. Claude Code, MCP servers, the prompt patterns that work, and what to delegate. Free.
Enter your work email. We send it straight over, plus a few short notes worth knowing. Unsubscribe any time.
