MCP Memory Service provides persistent memory for AI agent pipelines through an API and knowledge graph.
Open source
24 alternativesMemClaw
Caura, formerly MemClaw, provides governed shared memory for multi-agent fleets, with permissions and audit trails.
About MemClaw
Caura, formerly MemClaw, is an open-source (Apache 2.0) memory layer for AI agent fleets. Agents write plain text, and Caura stores it as searchable, governed memory intended for multi-agent and multi-tenant use, with MCP-native integration, trust tiers, keystone policies, audit trails, a knowledge graph, and self-improving retrieval. The project describes its loop as write, recall, compound.
A local quick start runs through Docker with Postgres, pgvector, Redis, an API and a scheduler. Standalone mode is single-tenant with authentication bypassed and needs no API key; without an AI provider key, memories are stored without embeddings and found by keyword, while adding a provider key enables semantic search. The repository notes that public agent-memory benchmarks such as LoCoMo and LongMemEval measure a single agent, single user and one long conversation, so fleet governance and workload-specific latency still need validation in the environment where they matter. Pricing is not stated in the source.
Key features
- Shared memory layer for multi-agent, multi-tenant AI fleets
- MCP-native integration with trust tiers and keystone policies
- Audit trails for stored agent memory
- Knowledge graph with self-improving retrieval over agent-written text
- Standalone local mode: single-tenant, authentication bypassed
- Keyword search without a provider key; semantic search with one
Best for: Teams building or running multi-agent AI fleets that need shared, policy-scoped memory with audit trails, and developers who want to try it locally first.
Good for
- Open source under Apache 2.0
- Existing memclaw_* tool calls and MEMCLAW_* environment aliases continue to work after the rename to Caura
- Local Docker quick start needs no API key or signup
Watch out for
- Benchmark coverage cited is single-agent, single-user, long-conversation; fleet governance and workload-specific latency still need validation
- Standalone mode bypasses authentication, and without an AI provider key memories are stored without embeddings and matched by keyword
- The PyPI memclaw-client package is only a redirect shell with no memclaw_client import or MemClaw class, and the npm MemClaw client was never published
MemClaw at a glance
- Category
- AI & Machine Learning › AI Agent Frameworks
- Type
- Open-source project
- Pricing
- Not confirmed. Check the site
- Open source
- Yes · Source code
- Language
- English
- Last checked
- Oct 6, 2026 · Online when checked
- Listed
- Oct 5, 2026
MemClaw alternatives
Tools and sites you could use instead, ranked by how closely they match.
SwarmClaw provides a self hosted runtime and multi agent framework with memory, tools, schedules, and delegation.
Open source
24 alternatives
CompozyOS coordinates agent command line tools as a team that shares project memory and tasks.
Open source
24 alternatives
PraisonAI supports agents that research, plan, code, and execute tasks, with memory and RAG features.
Open source
24 alternatives
Statewave provides reproducible memory context for agents using Postgres and pgvector.
Open source
24 alternatives
The open-multi-agent project provides a self-hosted TypeScript runtime with approvals and verifiable records.
Open source
24 alternativesPopular in AI Agent Frameworks
Well-known listings from the same topic.
Uplizd Unicornunicorn.uplizd.aiWebagentforge.netAbility AIability.aiHaystackhaystack.deepset.aiLangChainlangchain.comPromptShuttlepromptshuttle.com
MemClaw FAQ
What happened to the MemClaw name?
The product is now called Caura. Existing memclaw_* tool calls and supported MEMCLAW_* environment aliases still work, while new configuration should use caura_* names and the current Caura URLs.
Can I run Caura locally without an API key?
Yes. The Docker-based standalone mode runs single-tenant with authentication bypassed. Without an AI provider key, memories are stored without embeddings and found by keyword; a provider key adds semantic search.
What does the license allow?
The repository is published under Apache 2.0. No pricing is stated in the source.
Is something wrong or out of date?
Tell us and an editor will check it.
Own this website?
Claim this listing
Claim the listing and suggest updates yourself. An editor checks every claim first.