Codex support
Token Receipt ships a Codex skill that can be invoked from the thread with the exact workflow Codex already uses for skills. The skill calls the local runtime, reads the structured analysis, and then lets Codex write the final response in-session.
Codex prompt
Claude Code support
Claude Code gets the same skill-first flow: local parsing and image generation happen in the runtime, and Claude writes the final response using your existing session instead of a separate API key.
Claude Code prompt
What is Token Receipt?
Token Receipt is a screenshot-first usage artifact for coding agents.
It reads the session logs that Codex and Claude Code already write locally, turns those into structured usage facts, and lays them out as an itemized bill.
The result feels personal because it reflects your own habits: repeated file reads, subagent sprawl, context bloat, and every other expensive little ritual.
- Reads local logs from Codex and Claude Code
- Builds a thermal-paper PNG plus ready-to-edit post copy
- Uses skills so the host agent can write the final roast
Why this exists
The accurate observability product is not the point.
The point is a screenshot people instantly understand: your coding agent bill, officially itemized, with a line item for every bad habit you already know you have.
Real local signals make the output feel specific instead of generic.
What it itemizes
V1 focuses on signals we can defend from the logs:
Every bill is derived from local agent logs.
How it works
Token Receipt is local parsing plus agent-native copywriting, not a new hosted AI layer.
The runtime scans local Codex and Claude Code session logs
Deterministic heuristics turn those logs into receipt facts
The skill asks Codex or Claude Code to phrase the final response using the session you are already paying for
No prompt uploads by default. No telemetry in v1.
Who this is for
- People shipping with Codex or Claude Code every day
Open source
MIT licensed and open source.