> For the complete documentation index, see [llms.txt](https://blackprince001.gitbook.io/flowbench-documentation/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://blackprince001.gitbook.io/flowbench-documentation/getting-started/installation.md).

# Installation

FlowBench ships as two Go binaries and a Python package. The engine + CLI (`flowbench`) is all you need to author, run, and read tests; the agent (`flowbench-agent`) is copied onto the host under test when you want resource overlays; the Python SDK is optional, for teams authoring flows in Python.

## The engine + CLI

From a [GitHub release](https://github.com/blackprince001/flowbench/releases), download the archive for your platform, unpack it, and put `flowbench` on your `PATH`:

```bash
tar -xzf flowbench_<version>_<os>_<arch>.tar.gz
sudo mv flowbench /usr/local/bin/
flowbench version
```

Or build from source (Go 1.26+):

```bash
git clone https://github.com/blackprince001/flowbench.git
cd flowbench
go build -o flowbench ./cmd/flowbench
```

Every dependency is pure Go — there is no CGO, so cross-compiling with `GOOS`/`GOARCH` works without a toolchain for the target.

## The agent

The agent runs on the host under test, not the machine generating load. Download `flowbench-agent` from the same release (or `go build ./cmd/flowbench-agent`), copy it over, and start it:

```bash
flowbench-agent --addr :9090
```

Host sampling is implemented for **Linux only** in v1 — the agent reads `/proc`. It compiles everywhere, but on other platforms it serves an error and the run proceeds without an overlay (fail-open). See [the agent guide](/flowbench-documentation/guide/agent.md).

## The Python SDK

The SDK requires Python 3.10+ and depends only on `httpx`:

```bash
pip install flowbench
# or
uv add flowbench
```

To work against a checkout instead — developing the SDK itself, or tracking `main`:

```bash
pip install ./sdk-python
# or, working inside the repo with uv:
uv sync --project sdk-python
```

A release also attaches the wheel and sdist, which install the same way. The SDK shells out to the `flowbench` binary to resolve named targets, so install the CLI too — or pass `base_url=` to `flow.run()` and skip target resolution entirely. See [the Python SDK guide](/flowbench-documentation/guide/python-sdk.md).

## Verifying the install

```bash
flowbench version         # prints the build identity
flowbench help            # the four commands: run, serve, target, version
```

Next: the [Quickstart](/flowbench-documentation/getting-started/quickstart.md).


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://blackprince001.gitbook.io/flowbench-documentation/getting-started/installation.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.