Self-hosted long-term memory service that gives persistent AI identities continuity across clients, models, and agent interfaces.
oceanbase/powercontext
An open-source context service that keeps decisions, constraints, and work state outside chat sessions so agents and people can resume work across sessions.
- Stars
- 1,367
- Forks
- 305
- Open issues
- 143
- Last commit
- 30 minutes ago
What is it?
PowerContext is a server-side context service for AI agents. It stores durable work state such as decisions, constraints, evidence, and progress outside any single conversation, so a new session, a different agent, or another person can continue from the current state instead of rereading the transcript. It can run as a standalone Server with HTTP and MCP interfaces, or its core contracts can be embedded in another application.
The builtin runtime provides four artifact families: Memory for revisioned decisions and constraints, Handoff for recording work continuity between sessions and people, Experience for reusable lessons that pass through a review inbox, and Skill for reviewed procedures that must be exported explicitly before an agent can discover them. Every artifact is addressed by a family, stable ID, and revision, and reading an exact reference returns that historical snapshot.
Integrations cover Codex, Claude Code, OpenClaw, Hermes Agent, OpenCode, LangChain, LangGraph, Pydantic AI and others, with a Python client, CLI, and Dashboard. Memory writes work without a model API; automatic extraction from captured sources requires separate generation and embedding API credentials. Windows support is marked experimental.
Key features
- 01Stores decisions, constraints, and progress as revisioned artifacts outside chat sessions.
- 02Handoff records work continuity between sessions, agents, and people.
- 03Experience and Skill candidates pass through a review step before reuse.
- 04Recall is bounded by a byte budget and returns only active, approved content.
- 05Exposes HTTP, MCP, a Python async client, and an in-process Core SDK.
- 06Includes a setup wizard and a Dashboard for inspecting stored records.
Use cases
- Switching between coding agents such as Codex and Claude Code without losing project decisions.
- Handing an in-progress task to a teammate with a recorded current state and next action.
- Keeping long-lived project constraints available across many agent sessions.
- Building a custom agent application that needs server-side, inspectable memory.
Quick start
- 1
Install the PowerContext CLI and Server with uv
uv tool install --force "powercontext[cli,server]==1.4.0" - 2
Create a config folder and run the interactive setup wizard
powercontext config init --language en --output .env - 3
Start the local Server and keep the terminal open
powercontext server run --env-file .env
Things to know
- Windows support is experimental; macOS and Linux are the supported platforms.
- Full memory mode requires separate Generation and Embedding API credentials.
- Setup needs Python 3.11+, Git, and uv installed first.
In collections
Topics
FAQ
- What is powercontext?
- An open-source context service that keeps decisions, constraints, and work state outside chat sessions so agents and people can resume work across sessions. PowerContext is a server-side context service for AI agents. It stores durable work state such as decisions, constraints, evidence, and progress outside any single conversation, so a new session, a different agent, or another person can continue from the current state instead of rereading the transcript. It can run as a standalone Server with HTTP and MCP interfaces, or its core contracts can be embedded in another application.
- Is powercontext free to use? What license does it have?
- powercontext is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
- Is powercontext still maintained?
- Actively maintained. The last commit to powercontext was on Oct 22, 2026.
- What can I use powercontext for?
- Switching between coding agents such as Codex and Claude Code without losing project decisions. Handing an in-progress task to a teammate with a recorded current state and next action. Keeping long-lived project constraints available across many agent sessions. Building a custom agent application that needs server-side, inspectable memory.
- How do I install powercontext?
- Install the PowerContext CLI and Server with uv: uv tool install --force "powercontext[cli,server]==1.4.0" Create a config folder and run the interactive setup wizard: powercontext config init --language en --output .env Start the local Server and keep the terminal open: powercontext server run --env-file .env
- Where can I get powercontext?
- The source code and installation instructions are on GitHub: https://github.com/oceanbase/powercontext
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A Java framework for building agentic, workflow and multi-agent AI applications on top of Spring AI.
A local-first AI companion daemon for Windows and macOS that keeps long-term memory of you, turns chats into office files, and can extend itself.
A free, open-source Chinese-language book on AI Agent design principles and engineering practice, with 109 companion experiments and downloadable PDF/EPUB.
A memory layer for AI agents and apps that stores and retrieves user, session, and agent context across conversations.
A local-first, Markdown-native long-term memory layer for AI agents, storing memories as editable files with SQLite and LanceDB indexes.