- Updated capabilities for NVIDIA models to enforce OpenAI-compatible reasoning formats.
- Added new models: MiniMax M3, GLM 5.2, DeepSeek V4 Pro, DeepSeek V4 Flash, Kimi K2.6, and Nemotron 3 Ultra to the NVIDIA registry.
This enhances the provider's functionality and aligns with OpenAI standards.
Filter empty text blocks from thoughtSignature-only parts, preserve
tool_calls when functionResponse and functionCall coexist in the same
content, and skip empty regular text parts before they reach Claude.
Co-authored-by: Cursor <cursoragent@cursor.com>
Clients echo full message history each turn including reasoning_content,
which the Kimchi OpenAI gateway counts as input tokens. Multi-turn convos
balloon to 100k+ tokens and the model returns empty content.
KimchiExecutor.transformRequest now strips reasoning_content from assistant
messages when it exceeds an 8-char threshold, preserving the 1-char
placeholder injectReasoningContent sets and keeping content intact.
Co-authored-by: Cursor <cursoragent@cursor.com>
Gemini response translation wrote OpenAI-shaped bookkeeping into the
shared state.toolCalls map, which the downstream openai-to-claude
translator uses for Claude block metadata. That pre-population skipped
blockIndex creation, so Anthropic input_json_delta events lost index.
Track Gemini function calls via state.geminiToolCallCount instead,
leaving state.toolCalls clean for the Claude translator.
Closes#2218
Co-authored-by: Cursor <cursoragent@cursor.com>
Direct MITM server startup read rootCA.key/.crt immediately and exited
when either was missing, bypassing the manager.js CA setup path.
- generate Root CA from server.js when key/cert is missing
- make generateRootCA()/generateCert() synchronous to avoid a startup
race before readFileSync
- add unit test covering synchronous Root CA creation
Co-authored-by: Cursor <cursoragent@cursor.com>
Treat response.done as a terminal OpenAI Responses stream event so
passthrough streams ending with response.done are not flagged incomplete
and no synthetic response.failed is emitted. Restore the data: [DONE]
sentinel for same-format Responses passthrough streams.
Co-authored-by: Cursor <cursoragent@cursor.com>
Guard openai-responses compression: skip Headroom when body.input
contains non-message items (function_call, function_call_output,
reasoning) to preserve the Responses contract instead of collapsing
them into chat messages.
Co-authored-by: Cursor <cursoragent@cursor.com>
Opt-in quirk preserveCacheControl keeps cache_control on content blocks
for alicode/alicode-intl, enabling DashScope prompt caching. signature
is always stripped; all other providers unchanged.
Co-authored-by: Cursor <cursoragent@cursor.com>
CodeBuddy CN bonus packs ("Bonus Pack N") are one-shot credits whose
CycleEndTime equals DeductionEndTime — they expire for good and never
replenish. The dashboard rendered their resetAt as "Reset in Xd",
implying a monthly refill.
Tag bonus packs recurring:false (refill packs recurring:true) in the
usage handler, forward the flag through parseQuotaData, and word the
quota table / progress bar as "Expires in" / "Expires at" for
one-shot packs.
CodeWhisperer leaks literal <thinking> blocks into assistantResponseEvent,
duplicating reasoning already routed via reasoningContentEvent. Track
inThinking state to strip these tags during SSE transform, handling split
chunks across tag boundaries.
Co-authored-by: Cursor <cursoragent@cursor.com>
Generalize Claude 5h auto-ping into a provider-generic scheduler and add
opt-in Codex auto-ping that warms the next 5h window via a tiny gpt-5.5
request when session.resetAt slides. Default off, per-connection toggle,
failure cooldown, blocking-quota skip, drains stream before success.
Closes#2107
Co-authored-by: Cursor <cursoragent@cursor.com>
Use getModelInfo instead of parseModel in /api/translator/translate
so custom OpenAI/Anthropic-compatible provider prefixes resolve
correctly, aligning the debug path with the runtime chat path.
Fixes#1083
Co-authored-by: Cursor <cursoragent@cursor.com>
Cloudflare AI rule only sets flattenContent. Treat missing match as
provider-wide and missing drop as empty list to avoid crash. Fixes#1960.
Co-authored-by: Cursor <cursoragent@cursor.com>
Retry short-lived 5xx/capacity errors (500/502/503/504 + message
patterns) with bounded backoff capped at 15s; honor Retry-After/reset
hints and skip when wait is too long. Keep 400 non-retryable. Enable the
retry hook for 500 alongside existing 429/503.
Deduplicate sanitized Antigravity tool names before emitting the single
functionDeclarations group to avoid upstream "Tool names must be unique"
rejections.
Add Headroom size diagnostics and phantom-savings warning when reported
token delta does not shrink the outbound payload.
Co-authored-by: Cursor <cursoragent@cursor.com>
Carry Claude reasoning_effort/reasoning into OpenAI Chat, map into
OpenAI Responses reasoning.effort, and keep request-level effort
(incl. xhigh) across tool-result turns instead of collapsing to high.
Co-authored-by: Cursor <cursoragent@cursor.com>
Registry exposes the dashed id claude-opus-4-7; matchPattern treats "."
as a literal, so it missed the dotted pattern and fell through to the
generic claude opus entry (200k / claude-budget). Add an exact entry so
it resolves to 1M context + adaptive thinking, plus a unit test covering
the dashed Opus ids.
Co-authored-by: Cursor <cursoragent@cursor.com>
OpenAI-compatible apikey provider (chat/embedding/image) with dynamic
model discovery via modelsFetcher + passthroughModels. No executor needed.
Co-authored-by: Cursor <cursoragent@cursor.com>
Add 1M context capability overrides for claude-opus-4.8 and -thinking
variants, and use the resolved capability contextWindow (fallback 200k)
instead of the hardcoded 200k estimate in the Kiro executor.
Co-authored-by: Cursor <cursoragent@cursor.com>
Codex (openai-responses) body.input holds Responses items, not OpenAI
messages. Translate input -> OpenAI -> compress -> back to input so the
Responses contract is preserved. Fixes#1998
Co-authored-by: Cursor <cursoragent@cursor.com>
- mask API keys in usage stats/history responses (apiKeyMasked)
- validate proxy URL scheme + reject shell metachars before env write
- escape HTML in OAuth callback page to prevent XSS
- atomic O_EXCL lock file to prevent TOCTOU race in MITM startServer
- set mitmIsRestarting guard synchronously before any await
Co-authored-by: Cursor <cursoragent@cursor.com>
Forcing reasoning_effort:"medium" + reasoning_summary:"auto" on plain
requests tripped CodeBuddy's content filter and returned an error (#2071).
Make reasoning params opt-in: only set reasoning_summary when the client
sent an explicit reasoning_effort; none/off still drops it.
Fixes#2071
Co-authored-by: Cursor <cursoragent@cursor.com>
Pass Gemini AUDIO/TTS generateContent requests through to Google's native
v1beta endpoint instead of converting to chat, with per-credential fallback
(504 timeout, 502 fetch failure). Accept client keys from Bearer,
x-goog-api-key, or ?key= while forwarding only the configured Gemini
credential upstream. Expose native v1beta model names and rewrites, and add
Gemini 3.1 Flash TTS to the catalogs.
Co-authored-by: Cursor <cursoragent@cursor.com>
Add "text" to Codex Responses API allowlist so text.format.json_schema
reaches upstream for structured outputs instead of being stripped.
Co-authored-by: Cursor <cursoragent@cursor.com>
Keep provider-required streaming when client prefers JSON. The
Accept: application/json branch no longer flips stream back to false
for forceStream providers, fixing 400 errors on stream-only providers
(e.g. Command Code) for Hermes / Claude Code / other JSON clients.
Fixes#2031
Co-authored-by: Cursor <cursoragent@cursor.com>
Resolve Gemini TTS models from shared TTS catalog and provider registry
with a safe fallback, fixing requests resolving to models/undefined when
ttsConfig.models is empty. Add gemini-3.1-flash-tts-preview to catalogs.
Co-authored-by: Cursor <cursoragent@cursor.com>
Treat configured Headroom proxy as running when its /health endpoint
responds, even if local headroom CLI is not installed. Dashboard
Start/Stop stays limited to local loopback proxies while external
Docker sidecars can be enabled via HEADROOM_URL.
Closes#1948
Co-authored-by: Cursor <cursoragent@cursor.com>
Fixes#1933 — upstream returns 403 "Illegal access" on the chat endpoint
when requests lack a browser-like User-Agent. Mirror OmniRoute mimocode
executor: rotate across 3 Chrome UA strings on both bootstrap and chat.
Co-authored-by: Cursor <cursoragent@cursor.com>
codebuddy-cn shipped without a short alias, making its model prefix the
longest of any provider (codebuddy-cn/glm-5.2). Add alias + uiAlias
"cbcn" so the prefix becomes cbcn/glm-5.2, consistent with qoder (qd),
kiro (kr), codex (cx). "cbcn" = CodeBuddy CN; reserves "cbai" for a
future codebuddy-ai (intl) provider.
Backward-compatible: full id codebuddy-cn/... still resolves (resolver
matches both id and alias).
Also regenerate tests/__baseline__/alias-baseline.json: idToAlias entry
codebuddy-cn now maps to cbcn, and cbcn joins the sorted modelKeys list.
Co-authored-by: Cursor <cursoragent@cursor.com>
flattenToolHistory only recognized OpenAI-style tool calls. Anthropic-compatible clients (Claude Code, /v1/messages) send tool invocations and results as tool_use/tool_result blocks inside the message content array. Since panel calls strip the tools definitions (#1859), the panel expert models received structured tool history without schemas, causing them to fail or misbehave (leading to empty responses and 503 errors).
Extend flattenToolHistory to recognize and flatten tool_use and tool_result blocks in the content array into prose text, keeping panel expert execution robust and independent of the client API format.
PR decolua/9router#1910
Co-authored-by: Cursor <cursoragent@cursor.com>
Update OpenCode Go model catalog to match the official Go docs.
Route Qwen 3.7 and MiniMax models through /v1/messages with
Anthropic-compatible headers instead of OpenAI-compatible chat.
Keep GLM, Kimi, DeepSeek, and MiMo on /chat/completions; use
kimi-k2.7-code because the live Go API rejects kimi-k2.7 for chat
while the docs config example uses the code model ID.
PR #1931 by nguyenha935
Co-authored-by: Cursor <cursoragent@cursor.com>
Build on the optional Headroom Token Saver from Carmelo Campos
(PR: feat: add optional Headroom token saver). Add managed start/stop
of the local headroom proxy from the dashboard, install detection,
status probing, and a simplified Token Saver UI.
- detect headroom CLI + python>=3.10, probe proxy /health
- spawn/stop proxy as a detached, pid-tracked process
- /api/headroom/{status,start,stop} routes, gated local-only in dashboardGuard
- one-click Start/Stop Headroom modal, no manual config needed
- claude<->openai shape conversion for /v1/compress via 9router translators
Thanks to Carmelo Campos (@carmelogunsroses) for the original Headroom integration.
Co-authored-by: Cursor <cursoragent@cursor.com>
## Fixes
- **Kiro**: honor thinking effort budgets
- **AG/Kiro/Xiaomi**: provider fixes
- **Combo/Fusion**: flatten tool history in panel calls to prevent 503
- **LLM selector**: show custom vision models in selector and model list
- **Image**: prevent compatible nodes from shadowing provider aliases
Resolve the Kiro thinking budget from client intent (OpenAI reasoning_effort,
OpenAI Responses reasoning.effort, Claude output_config.effort, and Claude
thinking.budget_tokens) by reusing the shared thinkingUnified extractThinking
parser, then inject the resolved budget into the Kiro thinking system prefix.
Explicit none/off/disabled stops the prefix injection; synthetic -thinking
aliases keep the default budget.
Co-authored-by: Cursor <cursoragent@cursor.com>
Panel models in the fusion strategy must answer in prose. When the request
carried tools or prior tool_calls/tool messages, agentic panel models kept
emitting tool_calls instead of prose, so extractPanelText() returned empty
and the engine fell into the 503 "All fusion panel models failed" branch.
Panel fan-out now strips tools/tool_choice and flattens tool turns into
assistant prose (instead of dropping them), so panels keep the context but
cannot loop on tools. The judge still receives the unmodified history.
Co-authored-by: warelik <warelik@WARELIK-MB.local>
Co-authored-by: Cursor <cursoragent@cursor.com>
Expose user-added imageToText custom models as vision-capable chat
models in the default LLM selector and /v1/models, map custom service
kinds to runtime capabilities, and keep typed filtering for
/v1/models/{kind}.
Co-authored-by: Cursor <cursoragent@cursor.com>
Build a reserved prefix set from registry provider ids/aliases plus local
aliases, and skip compatible-node prefix matching for reserved prefixes so
built-in routes like cf/... stay on Cloudflare.
Co-authored-by: Cursor <cursoragent@cursor.com>
Adds Fusion as a third combo strategy alongside fallback/round-robin. A
fusion combo fans the prompt out to all member models in parallel, then a
configurable judge model synthesizes one final answer from the panel.
- handleFusionChat in open-sse/services/combo.js: quorum-grace collection
caps the straggler penalty, anonymized sources prevent judge brand-bias,
degrades to a direct answer on single survivor and 503 on total failure.
- chat.js dispatches strategy==="fusion" at both combo entry points.
- Combos dashboard: per-combo strategy Select replaces the round-robin
toggle, fusion reveals a judge picker, plus a strategy/capacity explainer.
- tests/unit/combo-fusion.test.js covers fan-out, judge routing/default,
quorum-grace straggler drop, single-survivor and total-failure degradation.
Co-authored-by: Cursor <cursoragent@cursor.com>
Adds long-lived API-key (ksk_) authentication for Kiro/AWS CodeWhisperer
and a direct claude:kiro / kiro:claude translation route that avoids the
lossy OpenAI two-hop pivot.
- translator: claude-to-kiro request + kiro-to-claude response translators,
registered on the exact source:target pair (direct route ahead of the
OpenAI pivot in index.js). claude-to-kiro uses shared schema constants
(ROLE/CLAUDE_BLOCK/DEFAULT_IMAGE_MIME) per app convention.
- auth: POST /api/oauth/kiro/api-key imports + validates a key via
ListAvailableProfiles, persists authMethod="api_key" (no refresh token).
- executor: send tokentype: API_KEY header and try *.amazonaws.com hosts
first for api-key creds; OAuth keeps kiro.dev first.
- fix: never inject the default placeholder profileArn for api-key auth
(CodeWhisperer 403s an ARN not owned by the key's account).
- ui: API Key method in the Kiro connect modal; surface api-key accounts
on the Quota Tracker and provider count.
- stream: env-overridable TTFT vs stall timeouts + Kiro keepalive frame.
- tests: claude-kiro-direct + kiro-profile-arn (11 tests).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Probes every chat model of every active provider with an image/audio content
block and classifies the outcome against declared capabilities. Surfaces models
where non-vision/non-audio models 400 on modality input (auto-strip candidates)
and where capability data is stale. Survey-only: logs a grouped table, never
fails on capability outcomes (cred/account noise filtered via status + message).
Co-authored-by: Cursor <cursoragent@cursor.com>
clientId/clientSecret của antigravity + gemini bị lặp 3 nơi
(registry, usage.js, src/lib/oauth). Gom vào shared.js
(ANTIGRAVITY_OAUTH_CLIENT, GOOGLE_OAUTH_CLIENT), các file spread vào.
Byte-for-byte: PROVIDERS/alias/oauth-url equal, golden 142 pass.
Thêm test guard nội dung + alias resolution.
Co-authored-by: Cursor <cursoragent@cursor.com>
Gộp switch 13 nhánh getUsageForProvider thành 1 registry object
(provider → handler), mỗi handler giữ nguyên signature/args qua ctx.
Behavior giữ nguyên (ollama vẫn chỉ nhận accessToken như cũ).
Thêm tests/unit/usage-dispatch.test.js guard dispatch.
Co-authored-by: Cursor <cursoragent@cursor.com>
Two `opencode` keys existed; the first (localhost:4096) was silently overridden
by the later one (opencode.ai, noAuth). Drop the dead entry. Resolved PROVIDERS
output unchanged (verified byte-for-byte).
Co-authored-by: Cursor <cursoragent@cursor.com>
CHAT_URL and opencode buildUrl base now read from PROVIDERS instead of repeating
the literal. Values identical; providers byte-for-byte + gate clean.
Co-authored-by: Cursor <cursoragent@cursor.com>
These translate-path builders had no runtime consumers: the translator route
uses executor.buildUrl/buildHeaders, and the barrel re-exports were unused.
Removing them eliminates the parallel URL/header build path (single source of
truth = executors). Drop their private helpers and the now-unused clineAuth
import. Golden executor snapshots + gate: no regression.
Co-authored-by: Cursor <cursoragent@cursor.com>
Snapshot buildProviderUrl/buildProviderHeaders/getTargetFormat for all providers
before merging the translate-path with executor URL/header builders.
Co-authored-by: Cursor <cursoragent@cursor.com>
Add helpers/usageHelper.js for conditional prompt/completion token details.
Apply to gemini/codex/claude response translators; keep each provider's token
math intact. No behavior change; golden + gate: no regression.
Co-authored-by: Cursor <cursoragent@cursor.com>
Add helpers/chunkBuilder.js; apply to claude/gemini/kiro/ollama/commandcode/
openai-responses response translators. Caller supplies id/created/model so each
keeps exact id-generation + usage semantics. Extend golden response stream to
openai-responses (codex). No behavior change; gate: no regression.
Co-authored-by: Cursor <cursoragent@cursor.com>
Lock current behavior before refactor: buildUrl/buildHeaders per default-executor
provider, translateResponse streaming (claude/gemini), translateRequest body
(openai->claude/gemini/kiro). Sanitize volatile fields (tokens, kimi device-id,
kiro conversationId, timestamps) for stable snapshots.
Co-authored-by: Cursor <cursoragent@cursor.com>
Bootstrap → JWT → chat flow with auto-retry. Inject the MiMoCode
system marker required to pass the upstream 403 anti-abuse gate,
derive JWT expiry from the exp claim, and use a stable per-machine
fingerprint as the bootstrap client.
Co-authored-by: Cursor <cursoragent@cursor.com>
convertOpenAIToolChoice mapped {type:"function"} verbatim and cloakClaudeTools
left tool_choice.name unsuffixed, both rejected by Claude on the cc/ path.
Map forced-function to {type:"tool",name}, allowlist Claude-valid types, and
suffix tool_choice.name when it targets a renamed client tool.
Fixes#1592
Co-authored-by: Cursor <cursoragent@cursor.com>
Added a new translation for the message "Endpoint is exposed without an API key." to various language files, enhancing user awareness regarding API security. This update ensures that users are informed about potential risks associated with unprotected endpoints in their respective languages.
MiniMax requires reasoning_content echoed back on assistant messages in
multi-turn/tool-call conversations. Add minimax and minimax-cn to
PROVIDER_RULES (scope all), same fix as DeepSeek (#1543).
Co-authored-by: Cursor <cursoragent@cursor.com>
Kiro requires a non-empty currentMessage tools array whenever history
references any tool use, else returns "Improperly formed request" (400).
Clients trip this by omitting tools on follow-ups after client-side
compaction.
- flattenToolInteractions(): no client tools -> collapse tool_use/result
to text so the "tools required" rule never fires
- reconcileOrphanedToolResults(): client tools -> salvage orphaned
results as text, keep matched ones, guard co-located tools array
- safeJSONParse(): guard tool-call argument parsing against bad JSON
- merge consecutive user userInputMessageContext; null-guard
currentMessage for assistant-only input
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Add shared OAuth credential lifecycle manager with provider-aware refresh
decisions. Implement CodexExecutor.refreshCredentials so 401/403 retry
refresh works for Codex, track lastRefreshAt and refresh before the
upstream stale-token window, preserve omitted idToken, and add
per-connection single-flight refresh to avoid refresh-token rotation races.
Merged from PR #1664.
Co-authored-by: Cursor <cursoragent@cursor.com>
MODEL_NO_MAP guard never re-routes Antigravity tab-autocomplete (tab_* models)
so latency-critical inline completion stays native. Flags gemini-3.5-flash-low
(agent/Default) as mandatory in the dashboard.
Co-authored-by: Cursor <cursoragent@cursor.com>
- Add qmodel_latest to QODER_MODEL_MAP
- Expose qmodel_latest in static Qoder provider catalog (qd)
- Generalize executor comment so model set does not go stale
- Add unit coverage for the new model key + catalog
Closes#1638
Co-authored-by: Cursor <cursoragent@cursor.com>
Kiro sends modelId "auto" for the main agent turn; without a defaultModels
slot getMappedModel returned null and the call leaked to AWS instead of the
configured provider. Adds the slot + guard test.
Co-authored-by: Cursor <cursoragent@cursor.com>
Route image model tests to /api/v1/images/generations and STT to
/api/v1/audio/transcriptions instead of forcing all non-embedding
models through chat completions. Adds kind-aware pingModelByKind,
hf->huggingface alias, and silent WAV sample for STT reachability.
Scoped to dashboard/internal model testing only; runtime inference
routing is unchanged.
Author: yicone <yicone@gmail.com>
Closes#1628
## Features
- Add Qoder provider: device-flow OAuth, COSY signing, WAF-bypass body encoding, live model catalog, dashboard quota tracker, 11 models (#1372)
- Add new models: Claude Opus 4.8 (Claude Code), GPT 5.4 Mini (Codex)
## Fixes
- DeepSeek thinking mode: echo `reasoning_content` back on follow-up/tool-call turns so OpenCode-free and custom providers no longer 400 with "reasoning_content must be passed back" (#1543)
- Reasoning injector: match deepseek/kimi model ids case-insensitively (covers custom providers using capitalized model names)
- OpenCode suggested-models: include free models without the `-free` suffix, e.g. `big-pickle` (#1535)
## Improvements
- Codex: trim sunset models, keep gpt-5.5 / gpt-5.4 / gpt-5.3-codex family, add gpt-5.4-mini
- volcengine-ark: refresh model list (add DeepSeek-V4-Flash/Pro, drop EOL entries)
- Lower stream stall timeout 35s → 30s for faster hang detection
Move device-flow / poll / userinfo / parseExpiry from src/lib/qoder/auth.js
into a QoderService class at src/lib/oauth/services/qoder.js, matching how
KiroService is organized. Also re-add the QoderService re-export from
services/index.js.
The split now mirrors Kiro:
src/lib/oauth/services/qoder.js OAuth flow (was auth.js)
src/lib/qoder/cosy.js Per-request signing (unchanged)
src/lib/qoder/encoding.js WAF-bypass body (unchanged)
src/lib/qoder/constants.js Endpoints + model map (unchanged)
Behavior is unchanged — same functions, same signatures, just relocated
into a class so the import path lines up with `import { QoderService } from
"@/lib/oauth/services"` like every other OAuth provider. parseExpiry is now
a static method so callers and tests can use it without instantiating.
42 tests still pass; build still clean.
Adds 18 new tests covering the bugs fixed in the previous commit so they
can't silently regress:
- parseExpiry (7 tests): numeric ms-epoch input, numeric strings handled
before Date.parse so "1700000000" doesn't get year-interpreted, RFC3339
strings, expires_in:0 honored as already-expired, 30-day fallback only
when both inputs are missing/invalid
- normalizeMessages (4 tests): system hoisting, multipart text flatten,
multiple system joining, empty input
- wrapQoderSSE (6 tests): the fixed cases — trailing partial line drained
in flush(), no chunks forwarded after [DONE], embedded newlines stripped
from inner body, error envelope produces error chunk + [DONE], non-ok
responses returned unchanged
- expose parseExpiry from auth.js, expose normalizeMessages/wrapQoderSSE
via __test__ from the executor (internals only — not part of the public
API). Marked with comment so the surface is intentional.
42 tests total (24 original + 18 new). Build still clean.
