OpenTrace started as an open-source graph explorer for repositories.
Paste in a codebase, generate a graph, explore relationships.
Useful — but temporary.
Today, we're launching the new app.opentrace.ai: a persistent workspace for understanding software systems. Save projects to the cloud, build multi-repository libraries, analyse pull requests in context, ask AI questions about your system, and share that understanding across your team — all from a single workspace that stays with you across sessions and machines.
From Temporary Exploration to Persistent Context
The old OpenTrace experience was session-based. You explored a repository, asked questions, inspected relationships — and when the session ended, the context disappeared with it.
The new platform changes that. Every indexed repository becomes part of a persistent project workspace: saved in the cloud, reopened instantly, expanded over time, and shared with collaborators.
Your Workspace Starts Immediately
Getting started is simple.
Open app.opentrace.ai and create a free account. Then paste in a GitHub or GitLab repository, connect a local directory, or import an existing graph.
OpenTrace immediately builds a map of how your codebase fits together, ready to navigate and explore.
The graph explorer is now the centre of the platform — no separate tools, no configuration required.
Local-First by Default
Your code never leaves your machine. OpenTrace builds everything locally — nothing gets uploaded to our servers.
Teams adopting AI-assisted workflows often can't centralise sensitive source code. OpenTrace gives you persistent, collaborative workspaces without that trade-off.
Indexing Is Visible
Indexing a large codebase takes time. Rather than a blank loading screen, OpenTrace shows you what it's doing as it goes — so you can watch the map take shape, not just wait for it.
(Fair warning: if you point it at a monorepo the size of a small country, your browser may stage a quiet protest. We're actively working on memory handling for very large repos — browser limits are a humbling adversary, and we respect them.)
Ask Questions About the System
The built-in AI assistant understands how your system fits together — it doesn't just search files, it understands structure.
You can ask:
- "Where is authentication handled?"
- "What calls this service?"
- "What depends on this component?"
- "What changed recently?"
- "What is the blast radius of this module?"
Every answer is grounded in the actual structure of the indexed system — not a fuzzy keyword match, not a guess.
You bring your own AI provider — Anthropic, OpenAI, Gemini, or a local model. Your credentials stay private and never touch our servers.
A direct integration with AI coding tools — Claude Code, Codex, Gemini, and others — is also coming soon, so you can query your OpenTrace projects without leaving your editor.
Pull Requests Become Structural
The new workspace adds persistent pull request analysis.
Instead of reviewing code changes in isolation, OpenTrace shows you what a pull request actually touches — which parts of the system it affects, and what might break downstream.
The challenge is no longer generating code — it's understanding the consequences of changes.
One Person Indexes. The Whole Team Benefits.
Projects inside OpenTrace are collaborative by design. Once saved, a repository can be shared across a team workspace — one engineer indexes, everyone else inherits the context. Collaborators stay synchronised, and updates propagate across shared projects.
Why This Matters
Software systems are getting too large and interconnected for shallow search to keep up. Teams spend time rebuilding context that should already exist — re-tracing dependencies, rediscovering decisions, mapping relationships that nobody documented.
OpenTrace holds that understanding persistently, so it doesn't have to be reconstructed every time.
Try It
Open app.opentrace.ai and create a free account. Index a repository, build a project, explore the graph, and ask questions about your system. Share the workspace with your team when you're ready.
Documentation and guides are at docs.opentrace.com.