Skip to content

Docs for Lite 2026.10.14, core 0.70.0 and ThinkWatch 3.5.0 - #70

Merged
fylorn merged 4 commits into
mainfrom
docs/2026-10-14
Oct 10, 2026
Merged

fylorn merged 4 commits into
mainfrom
docs/2026-10-14

Conversation

@fylorn

@fylorn fylorn commented Oct 10, 2026

Copy link
Copy Markdown
Contributor

Goes live with the releases; not to be merged before they are out.

  • Lite features page (en, zh-CN): the new upstream dialog's three groups of services (model vendors, platforms and relays, local and custom) with OpenRouter, ThinkWatch, Sub2API and New API / One API; the authentication methods each service has, with the fields a service decides filled in or left out; ChatGPT and Z.ai / BigModel sign-in inside the dialog, with a copyable sign-in link, ChatGPT's device code, and the device code only on a remote core; Edit upstream recognizing the service, and Z.ai / BigModel upstreams shown as an account sign-in with Sign in again; balances in the Quota / billing column with their sources, the enterprise gateway's key or account limits, the balance under Check connection and Refresh balance.
  • Claude Code, Codex and failover guides, CC Switch comparison, FAQ (en, zh-CN): the services to pick for GLM (Z.ai / BigModel with its Site, or the account sign-in) and for relays; DeepSeek and Kimi balances in the Claude Code guide; balances in the comparison's usage row.
  • Remote core guide, Lite overview and Lite page: the device code's address and code can be copied; upstreams show their balances.
  • Core docs from v0.70.0: the configuration reference gains an upstream's balance and signed_in; the server guide is unchanged.
  • Enterprise docs (en, zh-CN): GET /v1/usage in the gateway API reference (scope key or user, usage, limits with scope, kind, window, used and resets_at, expires_at, 401, 403 and 503) and in the deployment guide's endpoint table; a section on limits and budgets with route RPM and TPM caps now enforced and security.rate_limit_fail_closed; that setting in the configuration reference; the Limits tab of an API key's edit dialog; the ThinkWatch page's limits card. Deployment guide version examples move to 3.5.0.
  • Changelog: nothing to add here. The page lists every release of the three products from GitHub at build time, with Lite's Chinese notes from its annotated tags, so the three releases appear once they are published.

🤖 Generated with Claude Code

fylorn and others added 4 commits October 11, 2026 01:08
- Features (en, zh-CN): the new upstream dialog's three groups of
  services (model vendors, platforms and relays, local and custom) with
  OpenRouter, ThinkWatch, Sub2API and New API / One API; the
  authentication methods each service has, and the fields a service
  decides filled in or left out; ChatGPT and Z.ai / BigModel sign-in
  inside the dialog, with a copyable sign-in link, ChatGPT's device code,
  and the device code only on a remote core; editing that recognizes the
  service, and Z.ai / BigModel upstreams shown as an account sign-in; the
  Quota / billing column's balances, their sources and the enterprise
  gateway's key or account limits, the balance under Check connection
  and Refresh balance.
- Claude Code, Codex and failover guides, CC Switch comparison, FAQ
  (en, zh-CN): the services to pick for GLM, relays and the account
  sign-in; DeepSeek and Kimi balances; balances in the comparison.
- Remote core guide: the device code's address and code can be copied.
- Overview and Lite page: upstreams show their balances.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The configuration reference gains an upstream's balance and signed_in
keys; the server guide is unchanged since v0.67.1.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…be checked

- API reference (en, zh-CN): GET /v1/usage in the gateway endpoints,
  with what it answers about (the key, or its owner as a whole), usage,
  limits with scope, kind, window, used and resets_at, expires_at, and
  401, 403 and 503; a section on limits and budgets: what a request is
  checked against, the 429 and its label, route RPM and TPM caps now
  enforced, and security.rate_limit_fail_closed; the Limits tab of an
  API key's edit dialog and the limits endpoints it uses.
- Configuration (en, zh-CN): security.rate_limit_fail_closed in the
  security settings, covering Redis and database failures and route caps,
  with the console setting that holds it.
- Deployment guide (en, zh-CN): GET /v1/usage in the endpoints the
  console's Configuration Guide lists.
- ThinkWatch page: a client can ask what room its key has left, and the
  fail-open default covers the database as well as Redis.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@fylorn
fylorn merged commit 638fd4f into main Oct 10, 2026
1 check passed
@fylorn
fylorn deleted the docs/2026-10-14 branch October 10, 2026 17:22
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant