💡 Core Features

<div align="center">
  <a href="https://discord.acontext.io">
      <img alt="Show Acontext header banner" src="./assets/Acontext-header-banner.png">
  </a>
  <p>
    <h4>Context Data Platform for Building Cloud-native AI Agents</h4>
  </p>
  <p align="center">
    <a href="https://pypi.org/project/acontext/"><img src="https://img.shields.io/pypi/v/acontext.svg"></a>
    <a href="https://www.npmjs.com/package/@acontext/acontext"><img src="https://img.shields.io/npm/v/@acontext/acontext.svg?logo=npm&logoColor=fff&style=flat&labelColor=2C2C2C&color=28CF8D"></a>
    <a href="https://github.com/memodb-io/acontext/actions/workflows/core-test.yaml"><img src="https://github.com/memodb-io/acontext/actions/workflows/core-test.yaml/badge.svg"></a>
    <a href="https://github.com/memodb-io/acontext/actions/workflows/api-test.yaml"><img src="https://github.com/memodb-io/acontext/actions/workflows/api-test.yaml/badge.svg"></a>
    <a href="https://github.com/memodb-io/acontext/actions/workflows/cli-test.yaml"><img src="https://github.com/memodb-io/acontext/actions/workflows/cli-test.yaml/badge.svg"></a>
  </p>
  <p align="center">
    <a href="https://x.com/acontext_io"><img src="https://img.shields.io/twitter/follow/acontext_io?style=social" alt="Twitter Follow"></a>
    <a href="https://discord.acontext.io"><img src="https://img.shields.io/badge/dynamic/json?label=Acontext&style=flat&query=approximate_member_count&url=https%3A%2F%2Fdiscord.com%2Fapi%2Fv10%2Finvites%2FSG9xJcqVBu%3Fwith_counts%3Dtrue&logo=discord&logoColor=white&suffix=+members&color=36393f&labelColor=5765F2" alt="Acontext Discord"></a>
  </p>
  <div align="center">
    <!-- Keep these links. Translations will automatically update with the README. -->
    <a href="./readme/de/README.md">Deutsch</a> | 
    <a href="./readme/es/README.md">Español</a> | 
    <a href="./readme/fr/README.md">Français</a> | 
    <a href="./readme/ja/README.md">日本語</a> | 
    <a href="./readme/ko/README.md">한국어</a> | 
    <a href="./readme/pt/README.md">Português</a> | 
    <a href="./readme/ru/README.md">Русский</a> | 
    <a href="./readme/zh/README.md">中文</a>
  </div>
  <br/>
</div>


*Everyone is telling you how to use their agents. But what if YOU need to build an agent for 100,000 users, how would you start?*

**📦 Problem 1: 99% of your DB is just LLM messages.** 

> Poor schema design makes your most valuable data expensive and slow. Acontext handles context storage and retrieval via PG, Redis, and S3. 
>
> ChatGPT, Gemini, Anthropic, images, audio, files... we've got you covered.

**⏰ Problem 2: Long-running agents are a nightmare.** 

> You know context engineering, but you're always writing it from scratch. Acontext comes with built-in context editing methods and a todo agent out of the box.
>
> Managing agent state? Piece of cake.

**👀 Problem 3: You can't see how your agent is doing.** 

> How satisfied are your users, really? Acontext tracks tasks per session and shows you your agent's actual success rate. 
>
> Stop obsessing over token costs, improve the agent first.

**🧠 Problem 4: Your agent is hit or miss.**

> Can it learn from its wins? Acontext's experience agent remembers successful runs and turns them into reusable tool-use SOPs.
>
> Consistency is everything.



To solve those problems at once, Acontext becomes the **Context Data Platform**:

<div align="center">
    <picture>
      <img alt="Acontext Learning" src="./assets/acontext-components.jpg" width="100%">
    </picture>
  <p>Context Data Platform that Store, Observe and Learn</p>
</div>


# 💡 Core Features

- **Context Engineering**
  - [Session](https://docs.acontext.io/store/messages/multi-provider): unified message storage for any llm, any modal.
  - [Disk](https://docs.acontext.io/store/disk): save/download artifacts with file path.
  - [Context Editing](https://docs.acontext.io/store/editing) - manage your context window in one api.

<div align="center">
    <picture>
      <img alt="Acontext Learning" src="./assets/acontext-context-engineering.png" width="80%">
    </picture>
  <p>Context Engineering in Acontext</p>
</div>

- **Observe agent tasks and user feedback**
  - [Task](https://docs.acontext.io/observe/agent_tasks): collect agent's working status, progress and preferences in near real-time.
- **Agent self-learning**
  - [Experience](https://docs.acontext.io/learn/advance/experience-agent): let agent learn SOPs for each user.
- **View everything in one [dashboard](https://docs.acontext.io/observe/dashboard)**

<div align="center">
    <picture>
      <img alt="Dashboard" src="./docs/images/dashboard/BI.png" width="80%">
    </picture>
  <p>Dashboard of Agent Success Rate and Other Metrics</p>
</div>



# 🏗️ How it works?

<details>
<summary>click to open</summary>

```mermaid
graph TB
    subgraph "Client Layer"
        PY["pip install acontext"]
        TS["npm i @acontext/acontext"]
    end
    
    subgraph "Acontext Backend"
      subgraph " "
          API["API<br/>localhos

... (truncated)