Store, Observe, Learn

The context data platform for production agents

Multi-modal Storage·Task Monitoring·Pattern Learning

Identify successful execution patterns through the Store → Observe → Learn → Act loop, so agents act smarter and succeed more over time.

Cloud Platform

Open source

┌─[ACONTEXT]─────[Context Storage API]
# Create session with user association
session = client.sessions.create(user="alice@acme.com")
# Store messages - auto-converts between providers
client.sessions.store_message(
    session_id=session.id,
    blob={"role": "assistant", "content": "..."},
    format="openai" # openai | anthropic | gemini
)
# Upload files to S3-backed disk
client.disks.artifacts.upsert(
    disk.id, file=FileUpload(...),
    file_path="/docs/"
)
# Search with glob & regex patterns
client.disks.artifacts.glob_artifacts(
    disk.id, query="**/*.md")
💬
Sessions
Store and retrieve messages across OpenAI, Anthropic, and Gemini formats with automatic conversion.
OpenAIAnthropicGeminiMulti-modal
📦
Disk & Artifacts
S3-backed file storage with glob and regex search. Upload, download, and manage agent artifacts.
Glob searchRegex grepMetadata

Context Data Platform for Self-learning Agents

Everything you need to build, deploy, and scale production-ready AI agents

Sandbox Skills

Enable agents to practice and learn in isolated environments. Capture successful tool-calling patterns and transform them into reusable skills.

Context Engineering

Reduction, Compression, Offloading, Claude skills …… all in one backend

Artifact Disk

Filesystem-like workspace to store and share multi-modal outputs (.md, code, reports), ready for multiple agents collaboration.

Open Source

Architecture built on community standards. Every major dependency has a direct open-source alternative to ensure stack portability.

Multimodal Context Storage

Unified, persistent storage for all agent data, eliminating fragmented backends (DB, S3, Redis).

Background Observer

An intelligent scratchpad for long-running tasks, trace agent status and success rates in real-time.

SDKs & Integrations

Ready to use with OpenAI, Anthropic, LangGraph, Agno, and other popular agent frameworks

Acontext

User-Scoped · Direct Control · Self-Hostable

Upload skills per user, mount to sandbox, execute directly.

from acontext import AcontextClient
from acontext.agent import SANDBOX_TOOLS
from openai import OpenAI
 
client = AcontextClient()
llm = OpenAI(base_url="https://openrouter.ai/api/v1")
 
# 1. Upload skill with user identifier
skill = client.skills.create(
    file=("my_skill.zip", open("my_skill.zip", "rb")),
    user="alice@example.com"
)
 
# 2. Composable: Sandbox + Disk + Skills
sandbox = client.sandboxes.create()
disk = client.disks.create()
ctx = SANDBOX_TOOLS.format_context(client, sandbox.sandbox_id, disk.id, mount_skills=[skill.id])
 
# 3. Use ANY LLM via OpenRouter
response = llm.chat.completions.create(
    model="openai/gpt-4o",  # Or anthropic/claude-3.5-sonnet, deepseek/deepseek-r1...
    messages=[{"role": "system", "content": ctx.get_context_prompt()},
              {"role": "user", "content": "Run"}],
    tools=SANDBOX_TOOLS.to_openai_tool_schema()
)

Claude API

Workspace-Scoped · No Direct Control

Upload skills to workspace, use via container parameter only.

import anthropic
from anthropic.lib import files_from_dir
 
client = anthropic.Anthropic()
 
# 1. Upload skill (workspace-scoped only)
skill = client.beta.skills.create(
    display_title="My Skill",
    files=files_from_dir("/path/to/skill"),
    betas=["skills-2025-10-02"]
)
 
# 2. Use skill in container (no sandbox control)
response = client.beta.messages.create(
    model="claude-sonnet-4-5-20250929",
    max_tokens=4096,
    betas=["code-execution-2025-08-25", "skills-2025-10-02"],
    container={"skills": [{"type": "custom", "skill_id": skill.id, "version": "latest"}]},
    messages=[{"role": "user", "content": "Run"}],
    tools=[{"type": "code_execution_20250825", "name": "code_execution"}]
)

1 / 4 · Sandbox

Join the Community

Connect with early builders & preview new features

Join Discord

Star on GitHub