Supported Agents

The 10 coding agents and Hermes autonomous agent that TokenAnalytics tracks, how each is detected, and what data each captures.

TokenAnalytics reads log files from the agents listed below. Detection is automatic — no configuration needed. Just run your agent as normal and TokenAnalytics picks up the sessions.

Coding agents

Agent Vendor Log path Captures
Claude Code Anthropic ~/.claude/projects/ tokens, traces, reasoning, cost, subagents
Codex OpenAI ~/.codex/sessions/ tokens, traces, reasoning, cost
Gemini CLI Google ~/.gemini/ tokens, traces, cost
Antigravity Google ~/.gemini/antigravity/ traces, artifacts, screenshots, browser recordings
Qwen CLI Alibaba ~/.qwen/ tokens, traces
Vibe Local ~/.vibe/ tokens, traces, model
Cursor Cursor ~/.cursor/ + workspaceStorage tokens (estimated), traces, plans
GitHub Copilot GitHub VS Code chatSessions/ tokens (estimated), cost
OpenCode OpenCode ~/.local/share/opencode/ tokens, traces
Grok Build xAI ~/.grok/sessions/ tokens (estimated), traces, plans, cost

Autonomous agents

Agent Vendor Log path Captures
Hermes Agent Nous Research ~/.hermes/ tokens, traces, cost, subagents, skills, memory, cron jobs (read-only), source platforms

Hermes has a dedicated dashboard at /hermes — see Hermes Agent for details.

Detection notes

Claude Code

Reads JSONL session files from ~/.claude/projects/<dir>/. Captures input and output tokens, cost, tool calls, plan-mode output, reasoning blocks (with honest labelling if encrypted), and subagent delegation cards.

Codex

Reads session JSON from ~/.codex/sessions/. Captures tokens, tool calls, and reasoning steps including extended thinking when enabled.

Gemini CLI

Reads session files from ~/.gemini/. Captures token counts (input/output), cost, and the conversation trace. Note: Gemini CLI does not expose per-call cache-hit data.

Antigravity

Antigravity is Google's browser automation agent. TokenAnalytics reads its sessions from ~/.gemini/antigravity/ and additionally captures browser screenshots and Antigravity recording strips (thumbnail sequences from browser_recordings). See Artifacts.

Cursor

Cursor stores its chat sessions in ~/.cursor/ and VS Code's workspaceStorage/. TokenAnalytics merges sessions from both locations. Cursor's logs don't record token usage, so token counts are estimated from message length (Composer pricing) and flagged as estimates in the UI. Traces and plans (from Cursor's /plan output, in the Projects → Plans tab) are read directly.

GitHub Copilot

Reads Copilot's chat session files from VS Code's chatSessions/ directory (inside the VS Code user data folder). Output tokens and cost are read where the log includes them; input tokens are estimated from the prompt text. Copilot's logs don't expose a per-tool call list, so the trace shows the conversation rather than a tool-by-tool waterfall.

OpenCode

Reads session data from ~/.local/share/opencode/. Token counts and traces are captured; OpenCode children sessions are included in the relayed-session count (they are already counted in the parent session's token totals).

Hermes Agent

Hermes writes its state to ~/.hermes/ (or the directory pointed to by $HERMES_HOME). TokenAnalytics reads:

  • sessions.db / session JSON for per-call token and cost data
  • agent.log for per-API-call latency and cache-hit percentages
  • skills/, memory/, and soul/ directories for the configuration pages
  • Cron schedule files for the Schedules sub-page — read-only: TokenAnalytics shows Hermes's existing schedules and source platforms, but does not create, edit, or run them

The Schedules page is observe-only; manage cron jobs from Hermes itself.

Requesting a new agent

Don't see your agent? Open a GitHub issue with the agent name and its log file format and we'll add support.