Skip to main content

aiburn

Fast, fully local usage & cost reporting for Claude Code and Codex, in one table. See how much you're burning on your coding agents.

$ aiburn
aiburn · Claude Code + Codex usage

Date        Agent   Models       Input  Output  Cache    Cost
──────────  ──────  ───────────  ─────  ──────  ─────  ──────
2026-07-20  claude  opus-5       2.81K    222K  24.4M  $20.44
            codex   gpt-5.6-sol  64.0K   2.42K   500K   $0.64
2026-07-21  claude  opus-5         127   84.8K  22.3M  $17.27
──────────  ──────  ───────────  ─────  ──────  ─────  ──────
Date        Agent   Models       Input  Output  Cache    Cost
TOTAL                            66.9K    309K  47.2M  $38.35

Claude       $37.71   47.0M tokens
Codex         $0.64   567K tokens
Total        $38.35

Scanned 214 files in 61ms

Each raw model gets its own row, with repeated date and agent labels suppressed so model-level comparisons stay readable. Token counts are compact (K/M/B/T) and costs exact — pass --exact for full counts, or --json for raw integers. JSON groups mirror the table: one object per (date/month, agent, model), with the raw model name in models.

Why aiburn

  • Fully local & private — it only reads the log files and local Codex config your agents already write on your machine. No network calls or account; nothing about your usage ever leaves your computer.
  • One command, no install — uvx aiburn on macOS or Linux.
  • Both agents, one view — Claude Code and Codex side by side, by day, month, or session.
  • Fast at any size — scans gigabytes of history in well under a second, with near-instant startup, so it stays snappy as your logs grow.
  • Accurate cost — recomputed from token counts (including cache tiers), with date-aware model aliases and Codex Standard/Fast/Priority pricing.

Run it

No install needed — uv fetches a prebuilt binary and runs it:

uvx aiburn                 # macOS / Linux, any arch
uvx aiburn monthly

Keep it as a permanent command:

uv tool install aiburn     # then just `aiburn`

Or build from source with a Rust toolchain:

cargo install --git https://github.com/handasontam/aiburn

Usage

aiburn [command] [options]

Commands:
  daily      Per-day usage and cost (default)
  monthly    Per-month usage and cost
  session    Per-session usage and cost

Options:
  --since <YYYY-MM-DD>   Only include usage on/after this date
  --until <YYYY-MM-DD>   Only include usage on/before this date
  --claude               Only Claude Code
  --codex                Only Codex
  --all                  Show every session (session view; default: top 25)
  --exact                Full token counts instead of K/M/B/T
  --json                 Emit JSON instead of a table
  -h, --help             Show this help
  -v, --version          Show version

What it reads

Only your local coding-agent logs — read in place, never sent anywhere:

  • Claude Code — ~/.claude/projects (also ~/.config/claude and $CLAUDE_CONFIG_DIR)
  • Codex — ~/.codex/sessions and ~/.codex/archived_sessions, plus the top-level service_tier in ~/.codex/config.toml (or $CODEX_HOME)

A file that cannot be opened or read to the end is counted in the footer rather than silently skipped, so a permissions problem shows up as a warning instead of a smaller total.

Cost accuracy

Costs are recomputed from token counts with a built-in, pinned public pricing table, including Claude's 5-minute and 1-hour cache tiers and per-model cache rates (Fable 5.1 reads cache at $0.25/MTok, not the usual 10% of input). Claude Code figures match ccusage's to the cent; ccusage's headline total will still be higher if it also finds OpenCode, Gemini, or other agent logs that aiburn does not read. Codex logs carry no authoritative cost, so aiburn uses public list pricing: explicit priority/fast events use the matching Fast rate, missing tiers follow the top-level Codex config, and unknown tiers fall back to Standard. A Codex request whose whole context (fresh plus cached input) exceeds 272K tokens is billed entirely at OpenAI's long-context tier, as ccusage does.

Two details worth knowing:

  • Fast mode costs more per token. Fast/Priority usage appears as its own -fast row for both agents, priced at the per-model premium rate where one is published; otherwise at the base rate, which undercounts rather than dropping the row.
  • Codex replay history is skipped. Forking or resuming a session copies the parent rollout's records into the new file; those replayed token events belong to the parent's own file and are not charged again.

Acknowledgements

Inspired by ccusage by @ryoppippi and contributors — a great, broader tool that covers many more agents. aiburn is a smaller reimplementation focused on just Claude Code + Codex. ccusage is MIT licensed.

License

MIT

Metadata

Release files for aiburn 0.4.2

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Built distributions (wheels)

Table of built distributions (wheels) for aiburn 0.4.2
File Actions
aiburn-0.4.2-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl Python 3 none Linux glibc 2.17+ x86-64 Details for aiburn-0.4.2-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
aiburn-0.4.2-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl Python 3 none Linux glibc 2.17+ ARM64 Details for aiburn-0.4.2-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
aiburn-0.4.2-py3-none-macosx_11_0_arm64.whl Python 3 none macOS 11.0+ ARM64 Details for aiburn-0.4.2-py3-none-macosx_11_0_arm64.whl
aiburn-0.4.2-py3-none-macosx_10_12_x86_64.whl Python 3 none macOS 10.12+ x86-64 Details for aiburn-0.4.2-py3-none-macosx_10_12_x86_64.whl

Total release size: 1.1 MB

Release files / aiburn-0.4.2-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl

Download URL aiburn-0.4.2-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
Size 298.1 kB
Tags Linux glibc 2.17+ x86-64 Python 3
SHA-256 checksum
How to use checksums
76f3cd25a6d9c24ac49d8a2a128cbf689d2617080e6a60fc68ec14b50bfa23c6
BLAKE2b-256 checksum
How to use checksums
7e567d05d58705a2e989462edaf220cf546caf2fd5f5e9d8fb5507a82caa09fd
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 6, 2026.

Transparency log

Release files / aiburn-0.4.2-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl

Download URL aiburn-0.4.2-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
Size 282.3 kB
Tags Linux glibc 2.17+ ARM64 Python 3
SHA-256 checksum
How to use checksums
88c1594cd9d70c3a55638505f60646c7ee689664ffb298263143db1f4209b5e6
BLAKE2b-256 checksum
How to use checksums
03c3f76d6f602924601128de3fcca138347bf01e238688413786d1d5654a368d
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 6, 2026.

Transparency log

Release files / aiburn-0.4.2-py3-none-macosx_11_0_arm64.whl

Download URL aiburn-0.4.2-py3-none-macosx_11_0_arm64.whl
Size 256.2 kB
Tags Python 3 macOS 11.0+ ARM64
SHA-256 checksum
How to use checksums
5dad87adde10f582ede711b1ed62519fd71ce5bb98598e1f409fcaf710be5223
BLAKE2b-256 checksum
How to use checksums
fa54d21e51a511c6c96214eaa03f8dfbce6c02a1c4cba623c1cb0a82ec51c147
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 6, 2026.

Transparency log

Release files / aiburn-0.4.2-py3-none-macosx_10_12_x86_64.whl

Download URL aiburn-0.4.2-py3-none-macosx_10_12_x86_64.whl
Size 273.1 kB
Tags Python 3 macOS 10.12+ x86-64
SHA-256 checksum
How to use checksums
8f718478d30f4e782403c0f905c2c904420b8285fe495ba25e8f931b7a47fe6e
BLAKE2b-256 checksum
How to use checksums
9496cc5538a68d4fe727564f204c537881ca64aad39a45fcd064f2df5c2aa401
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 6, 2026.

Transparency log

Security for aiburn 0.4.2

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Full provenance

All 4 files in this release have been uploaded with provenance metadata.

Select from 4 files

Attestation for aiburn-0.4.2-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 6, 2026.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page