Our take
Activerecent releases- Activity
- 61
- Community
- 75
- Issues
- 46
- Pull requests
- 61
+441 stars in the last 30 days
- Releases more often than 81% of the projects we track
Memori is an actively maintained, multi-language memory layer with a broad deployment story, though teams should validate its integration and operating model against their own agent stack.
Good fit if
- You need agent conversations or execution to become structured, persistent state rather than relying on session context alone.
- Your stack uses supported Python or TypeScript integrations, or can connect through MCP, OpenClaw, or Hermes.
- You want a choice between managed service and private deployment, and can meet the runtime and infrastructure requirements.
Look elsewhere if
- You need a memory system tied to a specific LLM or datastore, rather than an LLM-agnostic layer.
- You need a turnkey deployment without Python or Docker-based infrastructure requirements.
- Your application does not need persistent agent memory or the integrations Memori provides.
All health signals
| Last commit | 2026-09-18 (1 month ago) |
|---|---|
| Commits, last 90 days | 5 |
| Releases, last 12 months | 32 (latest v3.3.6, 2026-05-28) |
| Contributors | 42 (top contributor: 29% of commits) |
| Issues closed, last 90 days | 1 |
| Pull requests merged, last 90 days | 5 (typically merged in 0 days) |
| Project age | 15 months |
Checked on 2026-10-19 with the GitHub API.
Project history
Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.
Stars
17,592Forks
3,998Change since Nov 16, 2026: +2 stars·+5 forks
Overview
Memori is LLM-agnostic memory infrastructure that turns what agents do, including conversations and execution, into structured, persistent state. It addresses session amnesia by capturing information for later recall instead of relying on an agent to repeat context or carry a large conversation history into every request.
Python and TypeScript SDKs integrate with existing LLM clients and frameworks, while integrations such as MCP, OpenClaw, and Hermes provide other ways to connect agents. Teams can use Memori Cloud or deploy in a single-tenant cloud, VPC, or on-premises environment, and the project also documents a bring-your-own-database option.
Key Features
- Captures conversations and agent execution as persistent memory for later recall.
- Adds attribution by entity and process, with sessions to group related interactions.
- Provides Python and TypeScript SDKs that integrate with supported LLM clients.
- Offers agent integrations including MCP, OpenClaw, and Hermes.
- Organizes augmented memory into attributes, events, facts, people, preferences, relationships, rules, and skills.
- Supports Memori Cloud and deployment choices including single-tenant cloud, VPC, and on-premises.
- Includes a dashboard for memories, analytics, playground use, and API keys.
Use Cases
- A product team can give a support agent persistent user context across separate conversations.
- Developers can connect coding agents through MCP so they can recall project conventions and preferences between sessions.
- Teams operating multi-step agents can group interactions into sessions and preserve execution context for later steps.
- Enterprise teams running production agents can evaluate private deployment options when agent workflows need persistent memory.
What you need
Detected in the repository
- Python >=3.10 (from pyproject.toml)
- A Dockerfile, so it can run in a container; a Compose file sets up the related services
- A test suite and automated checks on GitHub Actions
Getting Started
Install the Python SDK:
pip install memori
Set MEMORI_API_KEY and your LLM provider's API key, then register Memori with your client and set entity and process attribution. See the README for the TypeScript quickstart, configuration, and deployment documentation.
License in Plain Words
No standard license was detected. Without a license, the default copyright rules apply: you may read the code, but not reuse it. Check the repository or ask the authors.
Alternatives
- memoripy: Memoripy is a local Python runtime focused on evolving facts, scoped recall, and citations, rather than multi-language SDKs and managed deployment options.
- OpenViking: OpenViking organizes agent memories alongside knowledge and skills in a browsable context database, rather than focusing on conversation-to-state infrastructure.
- EverOS: EverOS stores memory as editable Markdown with local retrieval indexes, rather than offering cloud, VPC, and on-premises deployment modes.
| Project | Language | License | Stars | Status |
|---|---|---|---|---|
| Memori | Python | – | 17.2k | Active |
| memoripy | Python | Apache-2.0 | 694 | Active |
| OpenViking | Python | AGPL-3.0 | 39.2k | Active |
| EverOS | Python | Apache-2.0 | 13.4k | Active |
Considerations
- The project is active, with recent commits and frequent releases. It has dozens of contributors and is not concentrated around one person, but the measured recent issue-closure count is low, so support responsiveness is difficult to infer from that period alone.
- The Python package requires Python 3.10 or later. Docker and Docker Compose are listed among the project requirements; teams evaluating deployment should confirm which setup applies to their chosen mode.
- Memori's integrations and supported LLM clients are documented, but compatibility with a particular agent framework or production architecture should be tested before adoption.
- The README describes Apache 2.0 licensing; verify the repository's license file and distribution terms as part of your adoption review.
Found this useful?
Share it with someone who would like Memori.