Agent Flow is a local visualization tool for developers who want to understand what Claude Code and Codex do during a task. It runs in your browser or inside VS Code, Cursor, and Windsurf, turning agent activity into an interactive graph of tool calls, branches, and subagent coordination. The project is open source under Apache 2.0.
You can inspect individual agents and tool calls, follow their execution order, and spot slow calls or repeated work. A timeline and message transcript give you another way to review a session, while a file attention heatmap shows which files receive attention. The graph supports panning and zooming so you can examine larger runs.
Agent Flow detects Claude Code and Codex sessions concurrently and labels them by runtime. Tabs let you track separate sessions, and Codex activity includes reasoning and token counts drawn from its event stream. JSONL log support lets you replay recorded activity or watch events arrive in a log file.
The visualizer runs on your machine and reads local agent events. Its standalone app requires Node.js; the editor extension works in VS Code-compatible IDEs. The published standalone app sends anonymous usage telemetry by default, with an opt-out. That telemetry includes aggregate activity, model IDs, and system details, but excludes prompts, file paths, and tool-call contents. The VS Code extension and the app run from source send no usage telemetry.
Claim this page and we'll verify you by hand. Agent Flow gets the verified badge, and you can upgrade the listing to be featured on localhosted. Proud to be listed? Put our badge on your site.
Want more people to find Agent Flow?Promote it
Something wrong or outdated on this page?
779Updated 3 days agoMIT
macOS · Windows · Linux · Web#Git integration#LLM tracing#Multi-user access
Agent Console reads Claude Code and Codex session history on your own machines and shows usage in a local browser dashboard. It's for developers and teams who want to see where their coding agents spend tokens, how well caching works, and which sessions need attention. The open source project uses the MIT license and runs on macOS, Linux and Windows without an account.
758Updated 1 day agoMIT
macOS · Windows · Linux#LLM tracing#MCP
9.4KUpdated 4 hours agoApache-2.0
macOS · Windows · Linux · iOS · Android · Web#Batch processing#LLM tracing#Structured output
1KUpdated 4 days agoMIT
macOS · Windows · Linux · Docker · Web · VS Code#Agent Skills#LLM tracing#MCP
37.5KUpdated 5 days agoMIT
macOS · Windows · Linux · Docker · Web#Browser automation#LLM tracing#MCP
4KUpdated 6 days agoMIT
macOS · Windows · Linux · Docker · Web#LLM tracing#Works offline
agentacct is a local dashboard for developers who want to review what their coding agents did across projects and compare usage with recorded results. It brings task activity, file changes and test outcomes together with token counts and estimated costs. Claude Code and Codex have the fullest task records; it also supports OpenCode, Kimi Code and Hermes, with coverage varying by capability.
BAML is a programming language for developers building AI agents, with typed model calls and local tracing built into the language. It runs standalone on macOS, Linux and Windows, or alongside an existing application. The language is open source under Apache 2.0, and works offline.
Claude Code Agent Monitor puts Claude Code, Cursor and Codex activity in one local dashboard for developers tracking coding sessions and teams monitoring agent work. It runs in a browser served by your own machine, with desktop apps for macOS and Windows and a VS Code sidebar. The dashboard works offline, requires no account and sends no telemetry. It's free and open source under the MIT license.
Claude Code Router is an open-source local model gateway for developers who use coding agents and want to manage their model providers in one place. It runs on macOS, Windows and Linux, with Docker and a CLI with a browser interface also available. The project uses the MIT license.
Claude DevTools is a local debugging app for developers who want to inspect what Claude Code did beyond its terminal summaries. It reads existing session logs without modifying Claude Code, so you can review past work from terminal and IDE sessions. It's free and open source under the MIT license.