Practical overview
When it fits
Use 10xGraph when you are building an agent system and need to define its orchestration in Python, especially when persistence, tool coordination or an application-facing API is part of the task.
Who it is for
- Python teams building multi-agent workflows
- Developers connecting agents to model providers and MCP tools
- Teams that need checkpointed agent runs or a server around a compiled graph
- Python backend developers pairing an agent service with a TypeScript client
Documented uses
- Compose agents and tools into graph-based workflows.
- Connect workflows to OpenAI, Google Gemini or Anthropic models.
- Add MCP tools and parallel tool execution to agent systems.
- Checkpoint runs and pause a tool for human approval before resuming.
- Expose a compiled graph through the separately packaged REST API component.
Requirements & limits
The public repository identifies the project as open source under the MIT license. Core installation is through pip; the documentation requires Python 3.12 or newer.
- The core framework is Python; a TypeScript client is available as a separate ecosystem component.
- Provider SDKs, MCP, checkpointing backends and memory stores are optional integrations rather than all being required for the core package.
- Replay-safe completed tool calls require a checkpointer.
- The documented production checkpoint setup uses PostgreSQL with Redis; SQLite is described for local work.
- The REST API is provided by the separate 10xgraph-api package, not by the core graph definition alone.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Install the 10xGraph core and provider integrations
Set up the Python package with selected model-provider and MCP extras.
Before you start
- Python 3.12 or newer
- pip
- Provider credentials for any model provider you choose
Steps
Install the core package with the documented extras, for example: pip install "10xgraph[google-genai,openai,anthropic,mcp,pg_checkpoint]".
Set the environment variable for the provider you plan to use, such as GEMINI_API_KEY, OPENAI_API_KEY or ANTHROPIC_API_KEY.
If using PostgreSQL and Redis checkpointing, install the corresponding pg_checkpoint extra; other infrastructure integrations are optional.
Expected outcome
The core framework and selected provider or integration packages are installed and configured for development.
Guide limitations
- This setup does not create or run an agent graph or deploy an API server.
- The PostgreSQL-and-Redis checkpoint option requires those services; the README also describes SQLite for local work.
- Credentials and infrastructure must be supplied separately.
Compare options for Agent orchestration
These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.
| Project | Documented purpose | Platform | Access and requirements |
|---|---|---|---|
| 10Xgraph | Compose agents and tools into graph-based workflows. | CLI, Server, Web | The public repository identifies the project as open source under the MIT license. Core installation is through pip; the documentation requires Python 3.12 or newer. |
| Agenite | Define typed tools with JSON Schema validation and error handling. | CLI | |
| Veldra | Create an agent with a policy, tools, skills, knowledge base and reasoning method. | Self-hosted | The documented deployment is self-hosted and uses a sign-in gate after a first-run administrator account is created. |
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 14 October 2026
- Original website title
- 10xGraph/10xGraph · Official homepage title · retained in its source language