Ecosystem: ClawMetry — the Qwen Code reader is now free and open source (follow-up to #9294 / #9338)
Nobody has claimed this yet.
Assessment
- Difficulty
- 1/5
- Estimated time
- Under an hour
- Newbie friendliness
- 84/100
- Issue type
- Documentation
- Clarity
- Clearly specified
- Activity status
- Active
- Domain
- content, documentation
Research direction
Start at the official README's ## Ecosystem section and compare the existing entries' format. Done means adding the single ClawMetry repository link there, without a marketing-site link, then checking the rendered README and the surrounding list for formatting consistency.
Written by the indexing model from the issue text.
Description
Follow-up to #9294 and #9338, which @yiliang114 closed on 18 Sep on a single, clearly stated blocker. That blocker is now fixed, and the close said you'd be happy to reconsider if it were. Filing a fresh issue rather than commenting on the closed thread, since CONTRIBUTING asks for an issue before a PR and a comment on a closed issue is easy to miss.
Disclosure: I maintain ClawMetry, same as I disclosed in #9338.
What would you like to be added?
One line in ## Ecosystem, in the same format as the existing entries:
- [**ClawMetry**](https://github.com/vivekchand/clawmetry) — Local, zero-config observability for Qwen Code: sessions, transcripts, tool calls, model and token usage, read straight from `~/.qwen/projects/*/chats/*.jsonl`
Repo link only. #9338's second commit added a link to clawmetry.com and you were right that no existing entry carries a marketing-site link — that commit is gone and won't come back.
Why is this needed?
You closed on exactly one thing:
the Qwen Code reader is closed-source and paid even though the repo is MIT [...] we don't want the official README routing users to a capability behind a paywall. [...] Happy to reconsider if the Qwen Code adapter moves into the free, open-source package.
It moved, in vivekchand/clawmetry#6110 (merged 2026-09-20), released as clawmetry==0.12.890 on PyPI the same day (current: 0.12.892):
- The reader itself is now in the MIT repo:
clawmetry/adapters/qwen_code.py— 973 lines, 14 unit tests, two of them against a real, unmodified capture from@qwen-code/qwen-codev0.16.1 driven by a local Ollama. qwen_codeis inFREE_RUNTIMES, notPAID_RUNTIMES:clawmetry/entitlements.py.docs/ENTITLEMENTS.md— the document you quoted — now listsqwen_codeunder FREE and records why, citing this issue.
The "locked row" claim, checkable in a minute. Your concrete objection was that on the free tier pip install clawmetry && clawmetry renders a locked row for Qwen Code, so a reviewer can't reproduce the PR's verify steps. Here is that path with the entitlement gate forced on (CLAWMETRY_ENFORCE=1 — not the default grace mode, which would prove nothing), a clean HOME with no licence file, no cloud plan and no account:
$ python3 -m venv venv && ./venv/bin/pip install clawmetry # 0.12.892
$ HOME=/tmp/clean CLAWMETRY_ENFORCE=1 ./venv/bin/python - <<'PY'
from clawmetry.entitlements import get_entitlement
e = get_entitlement()
print("tier:", e.tier, "grace:", e.grace)
print("qwen_code ->", e.allows_runtime("qwen_code"))
print("claude_code->", e.allows_runtime("claude_code"))
from clawmetry.adapters.qwen_code import QwenCodeAdapter
a = QwenCodeAdapter()
print("detected:", a.detect().detected, "sessions:", len(a.list_sessions()))
PY
tier: oss grace: False
qwen_code -> True
claude_code-> False
detected: True sessions: 6
And from the dashboard's own entitlement API on that same clean install:
"tier": "oss", "grace": false,
"runtimes": ["goose", "nemoclaw", "openclaw", "qwen_code"],
"locked_runtimes": ["aider", "antigravity", "claude_code", "cline", "codex", "copilot", "cursor", ...]
qwen_code sits in the unlocked set; the paid runtimes are the locked ones. Starting it on that clean HOME prints [ok] Detected 1 agent runtime: Qwen Code, and the sessions come through with model, tool calls, thought reasoning parts and real usageMetadata token counts — e.g. a qwen3:8b session served by a local Ollama endpoint, where the token counts are the ones Qwen Code recorded and no dollar figure is invented for a model that has no public price. No account, no licence key, no paid wheel, no network call: the reader is local file reads only, and Qwen Code itself needs no change whatsoever.
This wasn't a carve-out for the listing. The rule in that entitlements doc, written when Goose moved free in August, is that an open-source runtime gets an open-source adapter while a commercial vendor product (Claude Code, Codex, Cursor, Copilot…) stays paid. Qwen Code is Apache-2.0, so the rule had already decided it — we just hadn't re-audited the list against it. Your close is what prompted the audit, and the doc says so.
Additional context
- On the earlier triage's Beta caveat: the adapter is no longer marked Beta, on the strength of the tests and the real capture above. If you'd prefer the entry to carry a qualifier anyway, propose the wording and I'll use it verbatim.
- Happy to reopen #9338 (one commit, README-only, disclosure in the body) whenever you say the word, or to leave it to a maintainer to add the line directly.
- If the answer is that you'd rather the official README not carry third-party dashboards at all, that's a completely fair call and I'll stop asking. I'm only back because the close named one fixable reason, and it's fixed.
Thanks for reading the entitlement split closely enough to catch it in the first place — that objection produced a better outcome than the listing would have.
- Dominant language
- TypeScript
- Stars
- 28k
- Forks
- 3.1k
- Avg merge
- 1d 1h
- Merged PRs (30d)
- 705
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from QwenLM/qwen-code
-
category/development priority/P3 scope/packaging scope/testing scope/web-shell type/bug
Difficulty 2/5 1-3 hours Newbie friendliness 86/100
-
Difficulty 1/5 1-3 hours Newbie friendliness 82/100
-
category/ui priority/P3 scope/settings scope/web-shell type/bug
Difficulty 2/5 Half a day Newbie friendliness 85/100
-
category/ui priority/P3 scope/settings type/feature-request
Difficulty 2/5 1-3 hours Newbie friendliness 88/100
-
priority/P3 scope/ci-cd status/waiting-for-feedback type/enhancement
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
All issues in QwenLM/qwen-code
Similar issues
-
calcite-components needs triage refactor
Difficulty 2/5 1-3 hours Newbie friendliness 75/100
Esri/calcite-design-system#15203 ·
-
Difficulty 2/5 1-3 hours Newbie friendliness 91/100
-
community first-timers-only good first issue hacktoberfest help wanted low hanging fruit up-for-grabs
Difficulty 1/5 Under an hour Newbie friendliness 95/100
-
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
Automattic/studio#4908 ·
-
Difficulty 2/5 1-3 hours Newbie friendliness 90/100