> For the complete documentation index, see [llms.txt](https://docs.visdom.virtuslab.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.visdom.virtuslab.com/visdom-tracing/visdom-ai-tracing.md).

# Visdom AI Tracing

The flight recorder for AI-assisted development. What the agent did, and proof of it.

Visdom AI Tracing records **what an AI coding agent actually did**, and can prove it afterwards. Every session, tool call, transcript, token count and file change is streamed off the developer's machine into a hash-chained, Ed25519-signed ledger on your server, with line-level attribution laid over git history.

It is deliberately not a reviewer and not a test gate. It does not judge whether the change is good; it records how the change was produced, and enforces the rules your organisation set for producing it.

Capture runs on **hooks, not a proxy**. `tracevault init` installs the coding agent's hooks plus two git hooks, and from that point the CLI streams the session as it happens and pushes commit metadata after every commit. Nothing has to be remembered, started or wrapped.

Two enforcement points sit on the same ledger:

* `tracevault check` runs on pre-push and can block a push that violated a policy
* `tracevault verify`, also shipped as a GitHub Action, fails CI when commits have no sealed trace on the server

Months later, that is what lets you answer, to somebody who was not there: which model, which prompt, which files, by whom, under which policy. The EU AI Act, SOX §404 and PCI-DSS questions get answered with records rather than recollection.

## Install

```bash
brew install VirtusLab/visdom-ai-tracing/tracevault
# or
cargo install tracevault-cli
```

The CLI is open source under Apache-2.0: [VirtusLab/visdom-ai-tracing-cli](https://github.com/VirtusLab/visdom-ai-tracing-cli).

![The dashboard: sessions, spend and policy findings across the repositories you have registered](https://2686717926-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F2dUajTQaSdFnN24ip7Qo%2Fuploads%2Fb5iEXurQxioxajzEiKnf%2Fdashboard.png?alt=media)

## Start here

* **Installing the CLI and signing in**: the binary, the device-grant sign-in, and API keys for CI
* **Setting up a repository**: what `tracevault init` installs and which agent it wires up
* **What a session captures**: the event stream, session statistics, and commit attribution
* **Command reference**: every command, and which ones are hooks

Deploying the server, the web UI and the API are documented at [tracevault.dev](https://tracevault.dev/docs/getting-started).

## The name on the binary

The component is **Visdom AI Tracing**. The CLI it ships is called `tracevault`, and so are its environment variables, its `.tracevault/` directory and the Homebrew tap, because the tool shipped under that name before the platform had one. Where this documentation says `tracevault`, it means the command; where it says Visdom AI Tracing, it means the component.

Within the Visdom platform, Context Fabric decides what an agent knows and the Orchestrator decides what it may run. Visdom AI Tracing is what makes both of them auditable.


---

# 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://docs.visdom.virtuslab.com/visdom-tracing/visdom-ai-tracing.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.