A stream that stalled or lost its upstream was closed with no terminal frame
at all, so clients saw "200 OK, a few chunks, then nothing" and could not tell
a truncated reply from a finished one. The Responses passthrough path already
synthesized response.failed; every other client format got nothing.
The watchdog now hands its reason ("stream stall timeout" or "upstream
connection lost") to onAbortTerminal, and buildStreamErrorBytes frames it per
client format: OpenAI-compatible clients get data: {"error":{...}} followed by
data: [DONE], Anthropic clients get `event: error`. The error frame always
precedes [DONE] (openai-python raises APIError on any data payload carrying an
error key), and no synthetic finish_reason is ever emitted — a truncated
stream must not look like a clean stop.
Co-Authored-By: Claude Code <noreply@anthropic.com>
codebuddy-intl: deepseek-v4-flash replaced by deepseek-v4.1-flash (same
gateway catalog as CN) and a capability override so the model keeps the
openai-style reasoning_effort path instead of the vendor-native "deepseek"
thinking shape the gateway rejects. Thinking levels low/high/xhigh.
ollama: add deepseek-v4.1-flash:cloud (verified on ollama.com/api/tags) with
vision + 1M context caps.
Co-Authored-By: Claude Code <noreply@anthropic.com>
The "Context window" dropdown wrote CLAUDE_CODE_MAX_CONTEXT_TOKENS, which
Claude Code ignores for any model it recognizes: its window resolver returns
the env value only when the id is unknown to the model table, so every
claude-* mapping kept the built-in 200K and the dropdown did nothing. It was
never the compaction threshold either.
- Replace it with CLAUDE_CODE_AUTO_COMPACT_WINDOW — the documented trigger
(100K–1M, clamped to the model window, env beats the autoCompactWindow
setting) — and relabel the field Auto-compact. The 1M preset becomes 700K,
which no longer collides with the marker it depends on.
- Add a "1M context" checkbox that appends the `[1m]` marker to the
ANTHROPIC_DEFAULT_*_MODEL envs. Claude Code assumes 200K unless the name
carries the marker — the resolver is a plain /\[1m\]/i test on the string,
so it applies to any id and no model lookup is involved; the user decides
which models are worth declaring as 1M.
- Toggling rewrites the model inputs immediately, and Apply writes them
verbatim, so a marker typed by hand is not stripped.
Rename maxContextTokens -> autoCompactWindow through the POST body and
RESET_ENV_KEYS so a reset clears the key actually written.
Co-Authored-By: Claude Code <noreply@anthropic.com>
Adds the Desktop-exclusive Preview models and the Xiaomi account-session
route to the existing xiaomi-mimo provider instead of a separate
xiaomi-desktop provider, so the dashboard shows one MiMo entry rather than
three overlapping ones.
Dual auth, same pattern as kimi — API key (sk-) covers the cloud API,
Desktop/OAuth adds the account session used by the Preview models:
- registry: category oauth, authModes [oauth, apikey], oauth block, the two
mimo-x-*-preview models, and the invite signupUrl
- executor: routes Preview models to the account-service route with a Cookie
session, everything else keeps the sourceFormat-matched transport
- oauth: custom ECDH encrypted-callback flow (X25519 -> SHA256 -> AES-256-GCM)
with a loopback callback proxy, plus one-click import of the local Desktop
auth.json
- usage: weekly quota from the account session
Fixes found while merging:
- the OAuth browser flow was dead: poll-status cleared the session before the
client could POST /exchange, so every exchange returned 400
- a Claude-format client was sent to /v1/chat/completions instead of the
declared /anthropic/v1/messages transport, because buildUrl ignored
runtimeTransport
- stopXiaomiMimoProxy leaked every pending session (each holding an X25519
private key) for the process lifetime
- the OAuth exchange did not persist the Desktop passToken, so the Preview
models could never work after a browser sign-in
Removes dead code: the local engine token minting (mimoEngine, never called
on the request path), the model-catalog and usage routes, engineToken/
engineUrl plumbing, and an unread top-level usage block.
Adds tests/unit/xiaomi-mimo-{executor,oauth-session,oauth-proxy}.test.js —
the provider previously had none.
Covers the 24 commits since the v0.5.69 tag. The package version already moved
to 0.5.75 in 4a390685b (CLI model selector), so the release commit is the
changelog alone, matching the convention in eb712ca82/4eda76e2a.
Co-Authored-By: Claude Code <noreply@anthropic.com>
qwen.js points at the same transport as alims-intl — identical baseUrl
(dashscope-intl compatible-mode), headers, and quirks — but carries only 8
Qwen models against alims-intl's broader catalog, so it adds nothing a user
could not already reach. Node count is back to 81.
The id was dropped on 2026-08-05 (dcdd4628b) when the Qwen OAuth flow died;
this removes the standalone API-key entry that shadowed it. The `qw` alias
returns to unassigned, its state before today.
Co-Authored-By: Claude Code <noreply@anthropic.com>
Registry order is the display order for the provider page, /v1/models, and the
CLI selector, so moving the entry to the head of `models` is the whole change.
deepseek-flash keeps its id and supportedFormats — only its position moves.
Co-Authored-By: Claude Code <noreply@anthropic.com>
A Vertex job id is a base64url-encoded operation name, and base64url decoding
accepts arbitrary bytes without throwing, so the previous decode-and-split
check let a crafted id splice a path traversal into the fetch URL while the
Bearer token stayed attached — e.g. "..%2F..%2Fevil" resolved to
/v1/evil:fetchPredictOperation on the Vertex host. body.model had the same
shape on the create path, where it is interpolated into the URL unescaped.
decodeJobId now requires a charset-only id, a byte-for-byte round-trip, and a
decoded name matching ^projects/{p}/locations/{l}/publishers/{pub}/models/{m}/operations/{op}$
— no field may contain "/", so ".." can never reach the URL. model ids are
restricted to [A-Za-z0-9._-].
Co-Authored-By: Claude Code <noreply@anthropic.com>
The merged Qoder work also rewrote shared translator/handler code so that
/v1/responses clients got token usage on response.completed. That changed
behaviour for every provider, not just Qoder: proxies saw input tokens
rise by the 2000-token context buffer, and the plain token mapping was
replaced by one that always adds input_tokens_details.
A probe confirms the Qoder benefit does not depend on those edits: the
executor's coalescer already emits one include_usage-style finish chunk, so
a Claude client receives input_tokens and cache_read_input_tokens with
every shared file at its original state. Only the Responses path relies on
the shared translator, and that path has no Qoder-owned seam to put it in.
Reverts the shared files to their pre-PR state and drops the Responses
usage test. The Cline envelope unwrap in nonStreamingHandler.js, which
landed after the PR in the same file, is kept.
defaultClaudeToolType() stamped tools[].type = "custom" onto every
Claude-format request carrying tools since e08ac6da. That satisfied
MiniMax (error 2013) but broke Anthropic-compatible endpoints that only
accept the legacy typeless tool shape. DeepSeek's endpoint
(api.deepseek.com/anthropic/v1/messages) whitelists its tool `type` enum
to the web_search_* variants and answers HTTP 400 "unknown variant
`custom`", so every Claude Code request routed to a DeepSeek connection
failed and surfaced as a persistent 503.
Run the defaulting only when the target provider declares the new
requireClaudeToolType quirk (MiniMax, MiniMax-CN). Add
shouldDefaultClaudeToolType(provider, finalFormat, tools, PROVIDERS) in
translator/concerns/toolCall.js so the gate is unit-testable, and cover
MiniMax keeping the explicit type, DeepSeek/Anthropic staying typeless,
non-Claude formats and tool-less requests never defaulting.
Effectively a no-op for MiniMax and a restore of the pre-e08ac6da
behaviour everywhere else. Another strict gateway now only needs the same
one-line quirk instead of a global behavioural change.
Anthropic accepts at most 4 blocks carrying cache_control per request. When the
client had already spent that budget, the re-anchor added a 5th marker and the
request was rejected with a non-retryable 400 that the failure path treated as
an account problem, retrying the same malformed body across the whole pool until
every account locked.
anchorClaudeCache now normalizes bare-object content, strips the invalid
cache_control carried by defer_loading tools, pins the 1h head anchors on the
last system block and last cacheable tool, then trims an over-budget body to 4
markers. The trim holds those head anchors and fills the remaining slots with the
tail-most message markers: a plain "keep the last four in document order" rule
drops the anchors first even though they lead document order, and skipping the
re-anchor at a spent budget left system/tools on the 5m default instead of 1h.
Some clients send content as a single block object rather than a one-element
array. Such a turn was dropped or zeroed on every leg that reads messages,
silently losing conversation history. normalizeMessageContent wraps it as a
one-block array on all four paths, and hasValidContent keeps it.
Cline (api.cline.bot) wraps non-stream chat completions in
{"success":true,"data":{...choices...}}, which both the dashboard model-test
ping and the proxy non-stream path read at top level, producing "Provider
returned no completion choices for this model" (#3644). Unwrap the envelope
before usage extraction and response translation; the error envelope
({"success":false,...}) never matches and passes through untouched.
Scoped through `transport.quirks.clineEnvelope` so only cline/clinepass opt
in — no other provider's response body is ever rewritten.
Also adds a live Cline catalog: `fetchClineRawModels()` is shared between
`resolveClineModels()` (full catalog, including free-tier ids such as
z-ai/glm-5.3-flash) and `resolveClinepassModels()` (cline-pass/* only), wired
into /v1/models, the per-provider models route, and the combo selector's
model picker with the static catalog kept as fallback.
Refreshes the dead api-airforce free models (anthropic/claude-3.7-sonnet,
moonshot/kimi-k2.6, google/gemini-2.5-flash) with the live gpt-oss-120b,
gpt-oss-20b and kimi-k2.7-code, plus passthroughModels, forceStream and a
suggested-models filter.
Cline/ClinePass requests failed with HTTP 401 ("Please make sure you are using
the latest version of Cline and re-authenticate your Cline account", #3230 /
#2333 / #3644). `getClineAccessToken()` unconditionally prefixed every token
with `workos:`, which is correct for Cline OAuth access tokens (WorkOS JWTs)
but wrong for ClinePass API keys — those are opaque strings (e.g. `clp_…`)
that the API accepts only verbatim, so the `workos:`-prefixed value was
rejected.
Only prefix tokens that look like a WorkOS JWT (`eyJ…`); API keys and other
opaque tokens pass through untouched, and an existing `workos:` prefix is
never doubled.
Also register `clinepass` in the token-refresh handlers. ClinePass shares
Cline's WorkOS auth endpoints, but without the entry expired ClinePass OAuth
tokens were never rotated, so every request kept 401ing. Finally, list
`apikey` first in the ClinePass `authModes` (ClinePass is meant to be used
with an API key from app.cline.bot/settings/api-keys), and add an "Import
from /models" button that pulls the live Cline catalog into custom models.
DeepSeek's Anthropic-compatible endpoint accepts only the built-in
web_search_20250305 / web_search_20260209 tools and rejects client-defined
`custom` tools (MCP / Read / Bash) with HTTP 400 "unknown variant `custom`".
The generic non-Claude filter in prepareClaudeRequest dropped the offending
tools but also dropped the web_search_* ones DeepSeek does accept.
- Add an opt-in per-provider transport quirk `claudeSupportedToolTypes`; when
declared it becomes a strict allow-list for Anthropic tool `type` values
- Stop stripping the `type` discriminator from surviving tools under that
quirk, since DeepSeek needs it to route built-ins
- Declare the quirk on the deepseek transport with the two web_search_* types
- Providers without the quirk keep the previous filter and normalisation
behaviour byte-for-byte; openai-format targets never reach this path
Codex's /responses validator has no Unicode property escapes, so a tool
`pattern` containing `\p{...}` 400s the whole request with `Invalid schema
for function ... is not a 'regex'` — identically on every account, costing a
full combo failover per turn (#3922).
- Add open-sse/utils/codexToolSchema.js: copy-on-write walk that drops only
`pattern` values carrying a property escape, returning the original
reference when nothing changed so the caller's schema stays intact for a
retry against another provider
- Treat `properties` keys as property names, so a field literally called
`pattern` is never read as the schema keyword; skip escaped literals via
backslash-parity counting
- Apply it in normalizeCodexTools for both function and namespace sub-tool
parameters, and log the strip count via dbg
- Add three cases to tests/unit/codex-tool-normalization.test.js
kiro.dev rejects any body carrying a top-level systemPrompt with
400 REQUEST_BODY_INVALID. The translators stopped emitting the field in
v0.5.59 (the prompt travels in the first user turn via contentPrefix),
but two paths kept writing it back downstream of the translator:
- rtk/systemInject.js::injectKiroSystem() appended the RTK prompt to
body.systemPrompt, so every kr/ model failed whenever an RTK injector
(caveman, ponytail) was active. It now appends to the first history
user turn's content (else currentMessage), reusing
dedupStringAppend/hasPrompt so retries stay idempotent.
- executors/kiro.js::appendRepairInstruction() wrote the tool-call repair
instruction to systemPrompt on the retry, turning every repair into a
hard failure. It now appends to currentMessage.userInputMessage.content.
isKiroBody() no longer requires a string body.systemPrompt — that marker
is gone from the wire shape — and sniffs the conversation turn shape
instead, keeping the stray-conversationState guard intact. Stale comments
in both kiro translators corrected: the systemPrompt local is only a
session-replay cache key, not a wire field.
Also drops the mirror/rollback repair heuristic the injector no longer
needs: net -52 lines.
Fixes#3641, #3845, #2890, #2901, #2939, #3109, #3459, #3749
- Coalesce Qoder's empty finish-in-delta frame with the later choices:[] usage
frame so OpenAI and Claude clients receive prompt_tokens, completion_tokens
and cache-hit tokens (the dashboard already saw them)
- Upload inlined images through /api/v2/image/upload like qodercli, and stub
oversized non-image files instead of stuffing 30MB+ data URIs into
agent_chat_generation
- Emit response.completed -> response.usage for chat-native upstreams so
/v1/responses clients (Codex CLI, sub2api) no longer log 0/0/0
- Keep Claude message_delta.usage working when usage arrives without choices[0]
- Escalate to the smallest advertised Qoder context tier (200K/400K/1M) when
the estimated prompt no longer fits max_input_tokens
- Pass apiKey for PAT connections and list hidden enable:false catalog keys
from /v1/models
- Add gpt-image-1.5, gpt-image-2, gpt-image-2.5, gpt-image-2.5-flare and
gpt-image-2.5-sunburst as Codex image models with multi-image support
- Add gpt-image-2.5, gpt-image-2.5-flare and gpt-image-2.5-sunburst to the
OpenAI provider catalog
- Route tool-backed image models through the Codex responses model while
passing the selected model to the image_generation tool, pinning
tool_choice and deriving generate/edit from the presence of references
- Cover the Codex gpt-image-2.5 request shape with a unit test
Video generation was xAI-only. Adds an adapter layer under
open-sse/handlers/videoProviders/ so /v1/videos/* can target OpenRouter or
Google Cloud credentials. A provider with no adapter keeps the exact previous
behaviour (raw body to {baseUrl}/{action}, poll {baseUrl}/{id}, verbatim
passthrough), so the xAI path is unchanged.
- openrouter: async job shape identical to xAI; creation POSTs to the /videos
collection root (no /generations suffix) and the registry HTTP-Referer /
X-Title headers are applied. Bodies pass through verbatim.
- vertex: two-way translation, since Veo does not speak the OpenAI-ish videos
shape. create -> :predictLongRunning { instances[], parameters{} }, poll ->
:fetchPredictOperation (Veo has no REST GET poll). The operation resource
name is base64url-encoded into the job id so GET /v1/videos/{id} stays a
flat path. Access tokens are minted from Service Account JSON via the
existing refreshVertexToken; raw API keys are rejected up front. The
operation response maps back onto the { id, status, video, videos } shape
clients already poll.
- videoCore: the request plan is rebuilt per attempt, so the 401 -> refresh
once -> retry once path picks up the refreshed token. Adapter validation
errors return 400 before any upstream call, so a malformed request can never
create a billable job.
- videoGeneration: GET /v1/videos/{id} resolves the provider from the pinned
x-connection-id connection, then ?provider=, then falls back to the xAI
default.
- registry: openrouter and vertex gain videoConfig, the video serviceKind and
video-kind models (Veo 3.1 / 3 / 2, Sora 2 Pro, Seedance 2.0).
The server's product-config payload (which the IDE plugin fetches from
copilot.tencent.com) publishes deepseek-v4.1-flash and no longer lists
deepseek-v4-flash, so the old id is dropped — same pattern as the previous
catalog refreshes (#3648, #3802). The v4-flash endpoint still answers 200,
but the published list is the contract.
Per the server table, maxOutput rises 50000 -> 128000 while contextWindow
stays 1000000.
- registry/codebuddy-cn.js: models[] entry swapped to the new id
- capabilities.js: per-model entry swapped, maxOutput -> 128000
No changes needed in thinkingLevels.js (the deepseek-v4* pattern already
matches the new id and publishes low/high/xhigh, matching the server's
supportedEfforts or pricing.js (the deepseek-v* glob yields the same rates).
EOF
)
Replace the flat numbered model list with provider-grouped browsing
(combos first, then providers by alias order), full-text search across
all models, and manual custom model ID entry. A single available
category opens directly into its model list.
Also bump root and cli packages to 0.5.75 and ignore packed
`9router-*` tarballs.
Co-Authored-By: Claude Code <noreply@anthropic.com>
The image handler's `version` header was commented out, so Codex image
requests reached chatgpt.com without the Version identity the backend
expects. Restore it and route every Codex identity header through one
constant.
The CLI version now lives on registry codex.transport as `cliVersion`
(the same pattern gemini-cli uses) and is re-exported as CODEX_CLI_VERSION,
so the registry User-Agent, the image handler and the connection test can
no longer drift apart. Bumped 0.136.0 -> 0.154.0 (current stable).
Co-Authored-By: Claude Code <noreply@anthropic.com>
Add the models the provider docs now list but the registry lacked:
chat/completions glm-5.3, kimi-k3, deepseek-flash, longcat-2.0,
hy4-preview, hy3
+ /messages qwen3.8-max, qwen3.8-flash
responses only grok-4.6, gpt-5.6-luna
Endpoints follow the table at https://opencode.ai/docs/go/. chat-only
models stay on the sourceFormat-matched transport guard so a Claude
client is never routed to /messages for a model that lacks it.
Co-Authored-By: Claude Code <noreply@anthropic.com>
Clear stale connection health state (modelLock_*, backoffLevel,
rateLimitedUntil, errorCode) whenever a connection is explicitly
marked active after successful validation or OAuth re-login.
Closes#3810
- Add muse-spark-1.2-contributor as responses-only model on OpenCode Go
- Normalize object tool schemas without properties in OpenCode Go executor
- Make fallback Responses call_ids unique across same-millisecond calls
- Make Responses output coercion fail-soft for circular and non-stringifiable values
- Recognize Fable weekly windows and normalize to weekly fable (7d)
- Fall back to 100% available weekly Fable window when Anthropic payload omits it
- Forward remaining percentages and enforce canonical Claude quota order in Quota Tracker
Co-Authored-By: Claude Code <noreply@anthropic.com>
- Group Antigravity Gemini text models into single 'Gemini (Flash / Pro)' quota
- Group Claude models into single 'Claude (Sonnet / Opus)' quota
- Prune stale or legacy model keys from hidden quota visibility list
Co-Authored-By: Claude Code <noreply@anthropic.com>
- Add muse-spark-1.3-contributor as responses-only model on OpenCode Go with dedicated executor
- Key Responses→chat streaming tool calls by item_id to prevent parallel tool calls merging into index 0
- Standardize tool coercions and call_id clamping in Responses API translation
Claude adaptive requests without an explicit effort are normalized to
output_config.effort: "high" instead of forwarding the unsupported
literal value "auto" which Anthropic rejects with HTTP 400.
- add a dedicated OpenCode Go executor that always sends x-opencode-session
- preserve a valid caller-provided native OpenCode session header
- translate downstream Agent session IDs into opaque, stable, Agent-scoped IDs
- forward the original provider session seed and client tool on both initial and credential-refresh requests
- Sync codebuddy-cn catalog and capabilities with copilot.tencent.com server payload
- Fix thinkingCanDisable semantics for glm-5.3 and deepseek-v4 models
- Add missing glm-5.2 thinking levels to thinkingLevels.js
- Add glm-5-turbo model to glm and glm-cn registries
- Registry/constants: drop qmodel_preview/gm51model, add lite,
qmodel_38max (Qwen3.8-Max), qfmodel (Qwen3.8-Flash), gmodel (GLM-5.3),
gfmodel (GLM-5.3-Flash)
- capabilities: add PROVIDER_CAPABILITIES['qoder'] so opaque internal
ids resolve to their real models' context windows and limits
- executor: preserve image blocks instead of flattening away, convert
Claude-style image blocks, and hash images into chat_record_id
- tests: cover image preservation, data-URI and Claude-block conversion
- build(docker): use CN mirrors for apk and npm
