Tracequest

Tracequest is a local web app and CLI for reading, searching, and analyzing coding-agent sessions.

TraceQuest is a local web application and command-line tool designed to help developers read and analyze coding-agent sessions. It displays messages, tool calls, errors, timing, and token usage from agents like Claude Code, Codex CLI, Cursor, Droid, OpenCode, Grok CLI, vktr, fx, and DeepSeek. Users can inspect activity timelines, token charts, and chapter breakdowns, as well as compare two sessions side by side. It also groups failed tool calls by cause, tracks estimated costs, and exports sessions as HTML files or Markdown text. The software runs entirely on the local machine without sending telemetry, supporting imports from remote machines over SSH and starting new agent runs using tmux.

Tracequest at a glance

Best for: Software engineer
Runs on: Web, CLI, self-hosted, open source
Pricing: Free
Released: October 2026
Category: AI coding tool
Website: github.com/viktor-com/tracequest

Key features

Search session messages and filter by project, source, machine, model, tool, and errors
Inspect activity timelines, token usage charts, and chapter-by-chapter prompts
Group failed tool calls by cause and view error rates, retries, and loops
Compare two sessions side by side to review duration, tokens, and outcomes
Read, start, and continue conversations with installed agents
Export sessions as single HTML files or Markdown snippets
Collect sessions from several machines over SSH using a hub configuration

Use it for

Find out why tool calls fail during coding agent sessions
Compare two coding sessions side by side for duration and token use
Search message text and filter agent sessions by project or error count
Export session reports as standalone HTML files or Markdown snippets
Collect and view coding sessions from multiple machines over SSH

Alternatives to Tracequest

Claude Tap — Inspect prompts, tool calls, and token usage locally without a hosted dashboard.
BugTape — Replays browser errors and failed requests for coding agents to fix automatically.
TraceRoot — Closes the self-improving loop for coding agents using production traces and evals.
Source Trace — Tracks code generated, kept, and rewritten to measure AI return on investment.
Tracecat — Builds agentic security workflows with open source control and custom MCP servers.
Agent Archive — Hand off coding sessions between Claude Code, Codex, and Cursor across machines.
Stroq — Scans recorded sessions and guards future actions with local hooks.
Opaline — Track team-wide token costs and coding session analytics.
Mindwalk — Visualize how coding agents explore and edit files on a 3D repository map.
Tracely — Turns production failures into regression tests that block pull requests.
Command Code — Repairs open model tool calls and learns developer patterns automatically.
Context Autopsy — Breaks down token usage by category and tracks processed tokens locally.

All Tracequest alternatives