MoonshotAI / MoonshotAI/kimi-code
Kimi Code login fails: HTTP 500 on /coding/v1 endpoints (me/models/chat) since 2026-08-29 18:05 CST
Nobody has claimed this yet.
- Dominant language
- TypeScript
- Stars
- 7.5k
- Forks
- 1.2k
- Avg merge
- 11h 53m
- Merged PRs (30d)
- 350
Description
Summary
kimi login fails with "Login failed: The server had an error while processing your request" — the Kimi Code coding API (api.kimi.com/coding/v1 and api.kimi.ai/coding/v1) has been returning HTTP 500 for all authenticated business endpoints since ~2026-08-29 18:05 CST (2+ hours and counting).
Repro
A freshly-issued OAuth token (device-code flow via auth.kimi.com succeeds; refresh token also succeeds) fails on every business endpoint:
curl -H "Authorization: Bearer <valid token>" https://api.kimi.com/coding/v1/me
# => HTTP 500
# {"error":{"message":"The server had an error while processing your request","type":"server_error"}}
curl -H "Authorization: Bearer <valid token>" https://api.kimi.com/coding/v1/models
# => HTTP 500 (same body)
curl -X POST -H "Authorization: Bearer <valid token>" -H "Content-Type: application/json" \
-d '{"model":"kimi-k2","messages":[{"role":"user","content":"hi"}]}' \
https://api.kimi.com/coding/v1/chat/completions
# => HTTP 500 (same body)
Same result on both regions (kimi.com and kimi.ai), both direct and via proxy.
Trace ID
The 500 responses carry Moonshot's own trace header:
x-trace-id: 1d802d265b22cb02756cb3fa50bd84a7
server-timing: inner; dur=172 (vs dur=4 for normal 404/401 responses)
The 404/401 responses come from the same nginx layer (x-trace-id present), so the request reaches the real backend; the 500 is produced after ~172ms of business processing.
What has been ruled out
- Not network: direct (residential CN egress) and proxy (HK datacenter) both 500
- Not DNS: 4 public resolvers (Ali/Tencent/Google/Cloudflare) all return the same volcddos CNAME
- Not MITM: TLS cert is a valid DigiCert
*.kimi.com - Not token: refresh grant works, freshly refreshed token also 500; replaced
device_idand re-authorized from scratch → still 500 - Not account: kimi.com web UI works fine — can log in and chat (web frontend is a different backend)
- Auth service healthy:
auth.kimi.comdevice authorization + token polling/refresh all return normal responses;api.moonshot.cn(platform API) returns normal 401
So the outage is isolated to the /coding/v1 business backend for Kimi Code.
Environment
- Ubuntu 26.04, kimi-code CLI 0.39.1 (both
--region mainland-cnand--region global)
Request
Please check the /coding/v1 backend (membership/plan validation DB is a common cause per the error reference). Happy to provide the full trace if needed.
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.
Research direction
Start by reproducing the failure with the provided curl requests against /coding/v1/me, /models, and /chat/completions, using the stated CLI version and both regions. No repository files or tests are named; done means determining whether the CLI has a code-level issue or the external coding backend is returning HTTP 500, with supporting trace details.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- typescript
- Domain
- api, backend, cli
- Issue type
- Bug
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Active
- Clarity
- Mostly clear
- Newbie friendliness
- 25/100