Installation

Install 10xGraph with pip install 10xgraph on Python 3.12 or newer, adding the extra for your model provider, then add 10xgraph-api for the server and CLI. Import the library as tenxgraph. This page also covers optional extras, API keys, verifying the install, and the TypeScript client.

What are the requirements?

Start with Python 3.12 or newer and a package manager like pip or uv. You also need an API key from the LLM provider you plan to use, and PostgreSQL and Redis if you deploy to production.

How do I install the Python packages?

Install 10xgraph (the core framework) and 10xgraph-api (the API server and CLI) in a Python virtual environment. A virtual environment ensures the 10xgraph command and the libraries come from the same Python interpreter, avoiding version conflicts.

Create the environment and activate it. With pip:

python3.12 -m venv .venv
source .venv/bin/activate  # on Windows: .venv\Scripts\activate
pip install 10xgraph 10xgraph-api

Or with uv (faster):

uv venv --python 3.12
source .venv/bin/activate  # on Windows: .venv\Scripts\activate
uv pip install 10xgraph 10xgraph-api

10xgraph is the core orchestration engine. 10xgraph-api adds a FastAPI server, a command-line tool, and a development playground.

Upgrading from the old names?

If you previously installed 10xscale-agentflow or 10xscale-agentflow-cli, uninstall them first because the new packages also ship an agentflow alias module:

pip uninstall 10xscale-agentflow-cli 10xscale-agentflow
pip install 10xgraph 10xgraph-api

Keeping both sets installed can leave you importing the wrong copy.

Which provider extra should I install?

10xGraph does not include LLM provider SDKs by default. You must install the extra for the provider you use. Choose one based on which model you want to call:

Install the extra in the same pip command or add it later:

pip install "10xgraph[openai]" 10xgraph-api

Combine extras with commas if you need more than one (here OpenAI, a Postgres and Redis checkpointer, and MCP tools):

pip install "10xgraph[openai,pg_checkpoint,mcp]" 10xgraph-api

Other optional extras

Beyond model providers, optional extras unlock advanced features:

The all extra installs every other extra at once and is useful for local development and CI, but pin specific versions before shipping to production.

Version pinning matters: 10xGraph is pre-1.0, so minor version updates may include breaking changes. Example of a safe lock:

pip install "10xgraph==0.10.0" "10xgraph-api==0.7.0"

Read the CHANGELOG.md before upgrading.

How do I verify the installation?

Run both commands to confirm everything is set up:

10xgraph version
python -c "import tenxgraph; print(tenxgraph.__file__)"

The first prints the 10xgraph-api version and the installed core framework version. The second confirms the Python library (tenxgraph) is importable from your virtual environment and shows where it is installed.

Example output (the exact formatting and versions vary by release):

$ 10xgraph version
10xgraph-api
  Version: 0.7.0
10xgraph (core)
  Version: 0.10.0

$ python -c "import tenxgraph; print(tenxgraph.__file__)"
/path/to/.venv/lib/python3.12/site-packages/tenxgraph/__init__.py

If the 10xgraph command is not found, make sure your virtual environment is activated. If the import fails, check that you ran pip install inside the activated environment. See installation troubleshooting for help.

How do I set LLM provider API keys?

Each provider reads its API key from an environment variable. You can export the variable in your shell session, or store it in a .env file that 10xgraph api loads at startup.

Export the key in your shell:

export OPENAI_API_KEY="sk-..."
export GEMINI_API_KEY="..."
export ANTHROPIC_API_KEY="sk-ant-..."

Or create a .env file in your project and point to it in 10xgraph.json:

{
  "agent": "graph.react:app",
  "env": ".env"
}

Then define the variables in .env (do not commit this file):

OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...

The variables 10xGraph recognizes are:

Never commit .env or API keys to version control. Use .gitignore to exclude it:

.env
.env.local

How do I install the TypeScript client?

If you are building a JavaScript or TypeScript application that calls the 10xGraph API server, install the client SDK:

npm install 10xgraph-client

The main class is TenxGraphClient, and the package requires Node.js 18 or newer. See the client documentation for usage.

The client talks to a running 10xgraph api server over HTTP, and you pass a thread ID to keep conversation history across runs. Use 10xgraph-api 0.7.0 or newer with client 0.6.0 or newer: the client authenticates WebSockets with the 10xgraph-bearer subprotocol, which older servers do not accept. HTTP calls work with either. See the Quickstart for a complete example.

What if something goes wrong?

If installation fails or the commands do not work, check these pages:

• Installation troubleshooting: pip errors, command not found, import errors, and environment variable issues
• Provider troubleshooting: model not found, auth failures, rate limit errors for your chosen provider
• API server troubleshooting: server startup errors and runtime issues
• Error code reference: what each structured error code means

Next step

Once installation is complete and verified, build your first agent in the Quickstart.