headroom
★ 77.3k
vs
kaeru
★ 40
headroom vs kaeru
headroom: Headroom compresses everything your AI agent reads — tool outputs, logs, RAG chunks, files, and conversation history — before it reaches the LLM. It achieves the same answers with a fraction of the tokens.; kaeru: Kaeru is a local-first, cross-agent cognitive engine for LLM agents, providing a typed property graph and a recollection layer for long-term memory. It enables multi-session and multi-agent continuity, allowing agents to maintain full context, follow provenance, and build on shared knowledge over time.
01
TL;DR
headroomMemory & Context · ★ 77.3k
When to choose:
- Long tool output, logs, JSON responses or RAG chunks are driving up cost or hurting answers.
- You want compression without code changes: run it as a local proxy, an MCP server, or by wrapping Claude Code, Codex or Cursor.
- Compression must run locally, with originals cached so the model can retrieve the full text.
When to skip:
- Your agent's problem is missing or outdated docs, not too much context: a docs server fits better.
- You cannot add a component that must keep running and may add latency.
- Your tasks cannot tolerate lossy compression: verify answer quality on your own work first.
kaeruMemory & Context · ★ 40
An agent needs to recall its thought process and project context from a previous session.
02
Side-by-Side Comparison
Field
headroomkaeruCategory
Memory & Context
Memory & Context
Stars
★ 77.3k
★ 40
License
Apache-2.0
MIT
Install Command
pip install "headroom-ai[all]"
git clone https://github.com/kaerulabs/kaeru.git && cd kaeru && cargo build --release
Updated
1d ago
2d ago
Open Source
Yes
Yes
Tags
Context Compression, Token Optimization, AI Agents
LLM Agents, Knowledge Graph, Bi-temporal Database
03
Features
headroom01In-app library for compression (Python/TypeScript)
02Zero-code-change proxy mode
03One-command agent wrapping for various AI agents
04Cross-agent shared memory and auto-deduplication
05Reversible compression (CCR) with original content retrieval
kaeru01Two-tier graph (operational/hippocampus and archival/cortex) for active thinking and settled knowledge.
02Bi-temporal nature with native assertion/retraction history for time-travel queries and non-destructive conflict resolution.
03Per-initiative scoping and layered re-entry to restore project context across sessions and access archival knowledge on demand.
04Optional cross-agent sharing through a cloud tier, allowing trusted teams to share settled knowledge with safety gates.
05Structural recall based on exact name lookup, typed graph traversal, and FTS fuzzy fallback, rather than primarily vector embeddings.
04
Use Cases
headroom↳Reduce LLM token usage and API costs for AI agents.
↳Enable shared context and memory across multiple AI agents.
↳Optimize coding agents by compressing tool outputs, logs, and RAG chunks.
↳Maintain full data fidelity with reversible context compression.
kaeru↳An agent needs to recall its thought process and project context from a previous session.
↳A team of agents needs to share and build upon a collective, long-term knowledge base.
↳An agent wants to capture new observations or hypotheses within a specific project scope.
↳Tracing the provenance and history of a piece of knowledge or a decision made by an agent.
05
Best For
headroomMost PopularEssential
kaeru—
FAQ
FAQ
What is the difference between headroom and kaeru?
Both headroom and kaeru are in the Memory & Context category. headroom has 77.3k stars, while kaeru has 40 stars.
Which is better, headroom or kaeru?
The best choice depends on your use case. Choose headroom if Reduce LLM token usage and API costs for AI agents., and kaeru if An agent needs to recall its thought process and project context from a previous session..
Is headroom free or open source?
Yes, headroom is open source on GitHub (Apache-2.0).
Is kaeru free or open source?
Yes, kaeru is open source on GitHub (MIT).
Cloud Infrastructure
AdDeploy on DigitalOcean
Host your MCP servers, Docker containers, and AI agents with $200 free credit over 60 days.
→