Skip to content

Never stop coding. Free MIT AI gateway: one endpoint, 375 providers (150+ free), 1,400+ models Kimi, Claude, GPT, Gemini, GLM, DeepSeek, MiniMax. Works with Claude Code, Codex, Cursor, OpenCode, Cline & Copilot. Quota-aware auto-fallback, RTK+Caveman compression saves 15-95% tokens, MCP/A2A, Desktop/PWA. Built by 900+ contributors

feat(api): aggregate combo model metadata in catalog (#2166)

Dohyun Jungdcb32a6ba06d
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 17c9ec1261..0ce5714d4c 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -138,48 +138,47 @@
Thank you to all **40 community contributors** who made v3.8.0 possible! 🎉
-| Contributor | PRs | Contributions |
-| :--------------------------------------------------------- | :-: | :------------------------------------------------------------ |
+| Contributor | PRs | Contributions |
+| :--------------------------------------------------------- | :-: | :------------------------------------------------------------------- |
| [@oyi77](https://github.com/oyi77) | 8 | #2010, #2014, #2041, #2052, #2061, #2074, #2091, #2094, #2096, #2131 |
-| [@backryun](https://github.com/backryun) | 6 | #1992, #2033, #2088, #2123, #2138, #2141 |
-| [@dhaern](https://github.com/dhaern) | 4 | #2028, #2039, #2087, #2090 |
-| [@Tentoxa](https://github.com/Tentoxa) | 2 | #2011, #2053 |
-| [@wauputr4](https://github.com/wauputr4) | 2 | #2009, #2046 |
-| [@payne0420](https://github.com/payne0420) | 1 | #2082 |
-| [@Tr0sT](https://github.com/Tr0sT) | 1 | #2012 |
-| [@AveryanAlex](https://github.com/AveryanAlex) | 1 | #2008 |
-| [@nickwizard](https://github.com/nickwizard) | 1 | #1991 |
-| [@rodrigogbbr-stack](https://github.com/rodrigogbbr-stack) | 1 | #1996 |
-| [@NekoMonci12](https://github.com/NekoMonci12) | 1 | #1999 |
-| [@congvc-dev](https://github.com/congvc-dev) | 1 | #2004 |
-| [@tatsster](https://github.com/tatsster) | 1 | #2007 |
-| [@xssdem](https://github.com/xssdem) | 1 | #2023 |
-| [@bypanghu](https://github.com/bypanghu) | 1 | #2027 |
-| [@herjarsa](https://github.com/herjarsa) | 2 | #2030, #2136 |
-| [@wucm667](https://github.com/wucm667) | 1 | #2031 |
-| [@hartmark](https://github.com/hartmark) | 2 | #2045, #2137 |
-| [@ddarkr](https://github.com/ddarkr) | 1 | #2047 |
-| [@tces1](https://github.com/tces1) | 1 | #2048 |
-| [@guanbear](https://github.com/guanbear) | 1 | #2054 |
-| [@Gi99lin](https://github.com/Gi99lin) | 1 | #2055 |
-| [@ivan-mezentsev](https://github.com/ivan-mezentsev) | 1 | #2063 |
-| [@JxnLexn](https://github.com/JxnLexn) | 1 | #2019 |
-| [@yoviarpauzi](https://github.com/yoviarpauzi) | 1 | #2092 |
-| [@rafacpti23](https://github.com/rafacpti23) | 1 | #2086 |
-| [@gleber](https://github.com/gleber) | 1 | #2103 |
-| [@rilham97](https://github.com/rilham97) | 1 | #2104 |
-| [@Gioxaa](https://github.com/Gioxaa) | 1 | #2105 |
-| [@boa-z](https://github.com/boa-z) | 1 | #2115 |
-| [@eleata](https://github.com/eleata) | 2 | #2116, #2133 |
-| [@rdself](https://github.com/rdself) | 1 | #2118 |
-| [@clousky2020](https://github.com/clousky2020) | 1 | #2119 |
-| [@abhinavjnu](https://github.com/abhinavjnu) | 1 | #2122 |
-| [@HoaPham98](https://github.com/HoaPham98) | 1 | #2089 |
-| [@christlau](https://github.com/christlau) | 1 | #2129 |
-| [@flyingmongoose](https://github.com/flyingmongoose) | 1 | #2134 |
-| [@05dunski](https://github.com/05dunski) | 1 | #1978 (cherry-picked) |
-| [@DavyMassoneto](https://github.com/DavyMassoneto) | 1 | #2140 |
-
+| [@backryun](https://github.com/backryun) | 6 | #1992, #2033, #2088, #2123, #2138, #2141 |
+| [@dhaern](https://github.com/dhaern) | 4 | #2028, #2039, #2087, #2090 |
+| [@Tentoxa](https://github.com/Tentoxa) | 2 | #2011, #2053 |
+| [@wauputr4](https://github.com/wauputr4) | 2 | #2009, #2046 |
+| [@payne0420](https://github.com/payne0420) | 1 | #2082 |
+| [@Tr0sT](https://github.com/Tr0sT) | 1 | #2012 |
+| [@AveryanAlex](https://github.com/AveryanAlex) | 1 | #2008 |
+| [@nickwizard](https://github.com/nickwizard) | 1 | #1991 |
+| [@rodrigogbbr-stack](https://github.com/rodrigogbbr-stack) | 1 | #1996 |
+| [@NekoMonci12](https://github.com/NekoMonci12) | 1 | #1999 |
+| [@congvc-dev](https://github.com/congvc-dev) | 1 | #2004 |
+| [@tatsster](https://github.com/tatsster) | 1 | #2007 |
+| [@xssdem](https://github.com/xssdem) | 1 | #2023 |
+| [@bypanghu](https://github.com/bypanghu) | 1 | #2027 |
+| [@herjarsa](https://github.com/herjarsa) | 2 | #2030, #2136 |
+| [@wucm667](https://github.com/wucm667) | 1 | #2031 |
+| [@hartmark](https://github.com/hartmark) | 2 | #2045, #2137 |
+| [@ddarkr](https://github.com/ddarkr) | 1 | #2047 |
+| [@tces1](https://github.com/tces1) | 1 | #2048 |
+| [@guanbear](https://github.com/guanbear) | 1 | #2054 |
+| [@Gi99lin](https://github.com/Gi99lin) | 1 | #2055 |
+| [@ivan-mezentsev](https://github.com/ivan-mezentsev) | 1 | #2063 |
+| [@JxnLexn](https://github.com/JxnLexn) | 1 | #2019 |
+| [@yoviarpauzi](https://github.com/yoviarpauzi) | 1 | #2092 |
+| [@rafacpti23](https://github.com/rafacpti23) | 1 | #2086 |
+| [@gleber](https://github.com/gleber) | 1 | #2103 |
+| [@rilham97](https://github.com/rilham97) | 1 | #2104 |
+| [@Gioxaa](https://github.com/Gioxaa) | 1 | #2105 |
+| [@boa-z](https://github.com/boa-z) | 1 | #2115 |
+| [@eleata](https://github.com/eleata) | 2 | #2116, #2133 |
+| [@rdself](https://github.com/rdself) | 1 | #2118 |
+| [@clousky2020](https://github.com/clousky2020) | 1 | #2119 |
+| [@abhinavjnu](https://github.com/abhinavjnu) | 1 | #2122 |
+| [@HoaPham98](https://github.com/HoaPham98) | 1 | #2089 |
+| [@christlau](https://github.com/christlau) | 1 | #2129 |
+| [@flyingmongoose](https://github.com/flyingmongoose) | 1 | #2134 |
+| [@05dunski](https://github.com/05dunski) | 1 | #1978 (cherry-picked) |
+| [@DavyMassoneto](https://github.com/DavyMassoneto) | 1 | #2140 |
## [3.7.9] — 2026-05-03
diff --git a/open-sse/config/providerRegistry.ts b/open-sse/config/providerRegistry.ts
index 8d55e95c57..7853ed71de 100644
--- a/open-sse/config/providerRegistry.ts
+++ b/open-sse/config/providerRegistry.ts
@@ -143,8 +143,13 @@ const KIMI_CODING_SHARED = {
"Anthropic-Version": ANTHROPIC_VERSION_HEADER,
},
models: [
- { id: "kimi-k2.6", name: "Kimi K2.6", contextLength: 262144, maxOutputTokens: 262144 },
- { id: "kimi-k2.6-thinking", name: "Kimi K2.6 Thinking", contextLength: 262144, maxOutputTokens: 262144 },
+ { id: "kimi-k2.6", name: "Kimi K2.6", contextLength: 262144, maxOutputTokens: 262144 },
+ {
+ id: "kimi-k2.6-thinking",
+ name: "Kimi K2.6 Thinking",
+ contextLength: 262144,
+ maxOutputTokens: 262144,
+ },
] as RegistryModel[],
} as const;
@@ -283,9 +288,9 @@ const CHAT_OPENAI_COMPAT_MODELS: Record<string, RegistryModel[]> = {
maritalk: buildModels(["sabia-4", "sabia-3.1", "sabiazinho-4", "sabiazinho-3"]),
"xiaomi-mimo": [
{ id: "mimo-v2.5-pro", name: "MiMo-V2.5-Pro", contextLength: 1048576, maxOutputTokens: 131072 },
- { id: "mimo-v2.5", name: "MiMo-V2.5", contextLength: 1048576, maxOutputTokens: 131072 },
- { id: "mimo-v2-omni", name: "MiMo-V2-Omni", contextLength: 262144, maxOutputTokens: 131072 },
- { id: "mimo-v2-flash", name: "MiMo-V2-Flash", contextLength: 262144, maxOutputTokens: 65536 },
+ { id: "mimo-v2.5", name: "MiMo-V2.5", contextLength: 1048576, maxOutputTokens: 131072 },
+ { id: "mimo-v2-omni", name: "MiMo-V2-Omni", contextLength: 262144, maxOutputTokens: 131072 },
+ { id: "mimo-v2-flash", name: "MiMo-V2-Flash", contextLength: 262144, maxOutputTokens: 65536 },
],
"inference-net": buildModels([
"meta-llama/Llama-3.3-70B-Instruct",
@@ -384,12 +389,48 @@ export const REGISTRY: Record<string, RegistryEntry> = {
tokenUrl: "https://console.anthropic.com/v1/oauth/token",
},
models: [
- { id: "claude-opus-4-7", name: "Claude Opus 4.7", supportsXHighEffort: true, contextLength: 1000000, maxOutputTokens: 128000 },
- { id: "claude-opus-4-6", name: "Claude Opus 4.6", supportsXHighEffort: false, contextLength: 1000000, maxOutputTokens: 128000 },
- { id: "claude-opus-4-5-20251101", name: "Claude Opus 4.5", supportsXHighEffort: false, contextLength: 200000, maxOutputTokens: 64000 },
- { id: "claude-sonnet-4-6", name: "Claude 4.6 Sonnet", supportsXHighEffort: false, contextLength: 200000, maxOutputTokens: 64000 },
- { id: "claude-sonnet-4-5-20250929", name: "Claude 4.5 Sonnet", supportsXHighEffort: false, contextLength: 200000, maxOutputTokens: 64000 },
- { id: "claude-haiku-4-5-20251001", name: "Claude 4.5 Haiku", supportsXHighEffort: false, contextLength: 200000, maxOutputTokens: 64000 },
+ {
+ id: "claude-opus-4-7",
+ name: "Claude Opus 4.7",
+ supportsXHighEffort: true,
+ contextLength: 1000000,
+ maxOutputTokens: 128000,
+ },
+ {
+ id: "claude-opus-4-6",
+ name: "Claude Opus 4.6",
+ supportsXHighEffort: false,
+ contextLength: 1000000,
+ maxOutputTokens: 128000,
+ },
+ {
+ id: "claude-opus-4-5-20251101",
+ name: "Claude Opus 4.5",
+ supportsXHighEffort: false,
+ contextLength: 200000,
+ maxOutputTokens: 64000,
+ },
+ {
+ id: "claude-sonnet-4-6",
+ name: "Claude 4.6 Sonnet",
+ supportsXHighEffort: false,
+ contextLength: 200000,
+ maxOutputTokens: 64000,
+ },
+ {
+ id: "claude-sonnet-4-5-20250929",
+ name: "Claude 4.5 Sonnet",
+ supportsXHighEffort: false,
+ contextLength: 200000,
+ maxOutputTokens: 64000,
+ },
+ {
+ id: "claude-haiku-4-5-20251001",
+ name: "Claude 4.5 Haiku",
+ supportsXHighEffort: false,
+ contextLength: 200000,
+ maxOutputTokens: 64000,
+ },
],
},
@@ -466,11 +507,41 @@ export const REGISTRY: Record<string, RegistryEntry> = {
// 1.05M). Public refs : openai/codex#19208, #19319, #19464 ;
// opencode#24171. max_output_tokens is stripped server-side
// (litellm#21193, codex#4138) so 128K is informational only.
- { id: "gpt-5.5", name: "GPT 5.5", ...GPT_5_5_CODEX_CAPABILITIES, contextLength: 400000, maxOutputTokens: 128000 },
- { id: "gpt-5.5-xhigh", name: "GPT 5.5 (xHigh)", ...GPT_5_5_CODEX_CAPABILITIES, contextLength: 400000, maxOutputTokens: 128000 },
- { id: "gpt-5.5-high", name: "GPT 5.5 (High)", ...GPT_5_5_CODEX_CAPABILITIES, contextLength: 400000, maxOutputTokens: 128000 },
- { id: "gpt-5.5-medium", name: "GPT 5.5 (Medium)", ...GPT_5_5_CODEX_CAPABILITIES, contextLength: 400000, maxOutputTokens: 128000 },
- { id: "gpt-5.5-low", name: "GPT 5.5 (Low)", ...GPT_5_5_CODEX_CAPABILITIES, contextLength: 400000, maxOutputTokens: 128000 },
+ {
+ id: "gpt-5.5",
+ name: "GPT 5.5",
+ ...GPT_5_5_CODEX_CAPABILITIES,
+ contextLength: 400000,
+ maxOutputTokens: 128000,
+ },
+ {
+ id: "gpt-5.5-xhigh",
+ name: "GPT 5.5 (xHigh)",
+ ...GPT_5_5_CODEX_CAPABILITIES,
+ contextLength: 400000,
+ maxOutputTokens: 128000,
+ },
+ {
+ id: "gpt-5.5-high",
+ name: "GPT 5.5 (High)",
+ ...GPT_5_5_CODEX_CAPABILITIES,
+ contextLength: 400000,
+ maxOutputTokens: 128000,
+ },
+ {
+ id: "gpt-5.5-medium",
+ name: "GPT 5.5 (Medium)",
+ ...GPT_5_5_CODEX_CAPABILITIES,
+ contextLength: 400000,
+ maxOutputTokens: 128000,
+ },
+ {
+ id: "gpt-5.5-low",
+ name: "GPT 5.5 (Low)",
+ ...GPT_5_5_CODEX_CAPABILITIES,
+ contextLength: 400000,
+ maxOutputTokens: 128000,
+ },
{ id: "gpt-5.4", name: "GPT 5.4", targetFormat: "openai-responses" },
{ id: "gpt-5.4-mini", name: "GPT 5.4 Mini", targetFormat: "openai-responses" },
{ id: "gpt-5.3-codex-spark", name: "GPT 5.3 Codex Spark" },
@@ -577,12 +648,48 @@ export const REGISTRY: Record<string, RegistryEntry> = {
{ id: "gpt-5.4-mini", name: "GPT-5.4 Mini", targetFormat: "openai-responses" },
{ id: "gpt-5.4", name: "GPT-5.4", targetFormat: "openai-responses" },
{ id: "gpt-5.5", name: "GPT-5.5", ...GPT_5_5_CODEX_CAPABILITIES },
- { id: "claude-haiku-4.5", name: "Claude Haiku 4.5", targetFormat: "openai-responses", contextLength: 200000, maxOutputTokens: 64000 },
- { id: "claude-sonnet-4.5", name: "Claude Sonnet 4.5", targetFormat: "openai-responses", contextLength: 200000, maxOutputTokens: 64000 },
- { id: "claude-sonnet-4.6", name: "Claude Sonnet 4.6", targetFormat: "openai-responses", contextLength: 200000, maxOutputTokens: 64000 },
- { id: "claude-opus-4-5-20251101", name: "Claude Opus 4.5 (Full ID)", targetFormat: "openai-responses", contextLength: 200000, maxOutputTokens: 64000 },
- { id: "claude-opus-4.6", name: "Claude Opus 4.6", targetFormat: "openai-responses", contextLength: 1000000, maxOutputTokens: 128000 },
- { id: "claude-opus-4.7", name: "Claude Opus 4.7", targetFormat: "openai-responses", contextLength: 1000000, maxOutputTokens: 128000 },
+ {
+ id: "claude-haiku-4.5",
+ name: "Claude Haiku 4.5",
+ targetFormat: "openai-responses",
+ contextLength: 200000,
+ maxOutputTokens: 64000,
+ },
+ {
+ id: "claude-sonnet-4.5",
+ name: "Claude Sonnet 4.5",
+ targetFormat: "openai-responses",
+ contextLength: 200000,
+ maxOutputTokens: 64000,
+ },
+ {
+ id: "claude-sonnet-4.6",
+ name: "Claude Sonnet 4.6",
+ targetFormat: "openai-responses",
+ contextLength: 200000,
+ maxOutputTokens: 64000,
+ },
+ {
+ id: "claude-opus-4-5-20251101",
+ name: "Claude Opus 4.5 (Full ID)",
+ targetFormat: "openai-responses",
+ contextLength: 200000,
+ maxOutputTokens: 64000,
+ },
+ {
+ id: "claude-opus-4.6",
+ name: "Claude Opus 4.6",
+ targetFormat: "openai-responses",
+ contextLength: 1000000,
+ maxOutputTokens: 128000,
+ },
+ {
+ id: "claude-opus-4.7",
+ name: "Claude Opus 4.7",
+ targetFormat: "openai-responses",
+ contextLength: 1000000,
+ maxOutputTokens: 128000,
+ },
{ id: "gemini-3.1-pro-preview", name: "Gemini 3.1 Pro", targetFormat: "openai-responses" },
{ id: "gemini-3-flash-preview", name: "Gemini 3 Flash", targetFormat: "openai-responses" },
{ id: "oswe-vscode-prime", name: "Raptor Mini", targetFormat: "openai-responses" },
@@ -606,12 +713,37 @@ export const REGISTRY: Record<string, RegistryEntry> = {
},
models: [
{ id: "auto-kiro", name: "Auto (Kiro picks best model)" },
- { id: "claude-opus-4.7", name: "Claude Opus 4.7", contextLength: 1000000, maxOutputTokens: 128000 },
- { id: "claude-opus-4.6", name: "Claude Opus 4.6", contextLength: 1000000, maxOutputTokens: 128000 },
- { id: "claude-sonnet-4.6", name: "Claude Sonnet 4.6", contextLength: 200000, maxOutputTokens: 64000 },
+ {
+ id: "claude-opus-4.7",
+ name: "Claude Opus 4.7",
+ contextLength: 1000000,
+ maxOutputTokens: 128000,
+ },
+ {
+ id: "claude-opus-4.6",
+ name: "Claude Opus 4.6",
+ contextLength: 1000000,
+ maxOutputTokens: 128000,
+ },
+ {
+ id: "claude-sonnet-4.6",
+ name: "Claude Sonnet 4.6",
+ contextLength: 200000,
+ maxOutputTokens: 64000,
+ },
// models for kiro free tier
- { id: "claude-sonnet-4.5", name: "Claude Sonnet 4.5", contextLength: 200000, maxOutputTokens: 64000 },
- { id: "claude-haiku-4.5", name: "Claude Haiku 4.5", contextLength: 200000, maxOutputTokens: 64000 },
+ {
+ id: "claude-sonnet-4.5",
+ name: "Claude Sonnet 4.5",
+ contextLength: 200000,
+ maxOutputTokens: 64000,
+ },
+ {
+ id: "claude-haiku-4.5",
+ name: "Claude Haiku 4.5",
+ contextLength: 200000,
+ maxOutputTokens: 64000,
+ },
{ id: "deepseek-3.2", name: "DeepSeek V3.2" },
{ id: "minimax-m2.5", name: "MiniMax M2.5" },
{ id: "minimax-m2.1", name: "MiniMax M2.1" },
diff --git a/scripts/scratch/check_usage.js b/scripts/scratch/check_usage.js
index 15e6cab603..9360bab904 100644
--- a/scripts/scratch/check_usage.js
+++ b/scripts/scratch/check_usage.js
@@ -1,34 +1,41 @@
-import Database from 'better-sqlite3';
-import path from 'path';
-import os from 'os';
+import Database from "better-sqlite3";
+import path from "path";
+import os from "os";
-const SQLITE_FILE = path.join(process.cwd(), 'data', 'storage.sqlite');
+const SQLITE_FILE = path.join(process.cwd(), "data", "storage.sqlite");
-console.log('Checking database at:', SQLITE_FILE);
+console.log("Checking database at:", SQLITE_FILE);
const db = new Database(SQLITE_FILE, { readonly: true });
try {
- const rows = db.prepare(`
+ const rows = db
+ .prepare(
+ `
SELECT api_key_id, api_key_name, COUNT(*) as count, MAX(timestamp) as last_used
FROM usage_history
GROUP BY api_key_id, api_key_name
ORDER BY count DESC
- `).all();
+ `
+ )
+ .all();
- console.log('Top Usage Entries:');
- console.table(rows);
+ console.log("Top Usage Entries:");
+ console.table(rows);
- const keys = db.prepare(`
+ const keys = db
+ .prepare(
+ `
SELECT id, name, key_prefix, machine_id
FROM api_keys
- `).all();
-
- console.log('All API Keys:');
- console.table(keys);
+ `
+ )
+ .all();
+ console.log("All API Keys:");
+ console.table(keys);
} catch (err) {
- console.error('Error:', err.message);
+ console.error("Error:", err.message);
} finally {
- db.close();
+ db.close();
}
diff --git a/src/app/api/v1/models/catalog.ts b/src/app/api/v1/models/catalog.ts
index 3d3f94cd77..5ffcfaf5f9 100644
--- a/src/app/api/v1/models/catalog.ts
+++ b/src/app/api/v1/models/catalog.ts
@@ -17,14 +17,18 @@ import { getAllVideoModels } from "@omniroute/open-sse/config/videoRegistry";
import { getAllMusicModels } from "@omniroute/open-sse/config/musicRegistry";
import { REGISTRY } from "@omniroute/open-sse/config/providerRegistry";
import { CODEX_NATIVE_UNPREFIXED_MODELS } from "@omniroute/open-sse/services/model";
+import { resolveNestedComboTargets } from "@omniroute/open-sse/services/combo";
import { getAllSyncedAvailableModels } from "@/lib/db/models";
import { getCompatibleFallbackModels } from "@/lib/providers/managedAvailableModels";
import { hasEligibleConnectionForModel } from "@/domain/connectionModelRules";
import {
INTERNAL_PROXY_ERROR,
enrichCatalogModelEntry,
+ getCanonicalModelMetadata,
getCatalogDiagnosticsHeaders,
} from "@/lib/modelMetadataRegistry";
+import { getSyncedCapability } from "@/lib/modelsDevSync";
+import { getModelSpec } from "@/shared/constants/modelSpecs";
import { isAuthRequired, isDashboardSessionAuthenticated } from "@/shared/utils/apiAuth";
import { parseModel } from "@omniroute/open-sse/services/model";
import { getTokenLimit } from "@omniroute/open-sse/services/contextManager";
@@ -54,6 +58,50 @@ const FALLBACK_ALIAS_TO_PROVIDER = {
qw: "qwen",
};
+type ComboCatalogTarget = {
+ modelStr?: string;
+ provider?: string | null;
+};
+
+type ComboTargetCatalogMetadata = {
+ contextLength?: number;
+ maxInputTokens?: number;
+ maxOutputTokens?: number;
+ inputModalities?: string[];
+ outputModalities?: string[];
+ capabilities: Record<string, boolean>;
+};
+
+function isPositiveFiniteNumber(value: unknown): value is number {
+ return typeof value === "number" && Number.isFinite(value) && value > 0;
+}
+
+function parseJsonStringArray(value: unknown): string[] {
+ if (typeof value !== "string" || value.trim().length === 0) return [];
+ try {
+ const parsed = JSON.parse(value);
+ return Array.isArray(parsed)
+ ? parsed.filter((entry): entry is string => typeof entry === "string" && entry.length > 0)
+ : [];
+ } catch {
+ return [];
+ }
+}
+
+function intersectStringArrays(arrays: string[][]): string[] {
+ if (arrays.length === 0 || arrays.some((values) => values.length === 0)) return [];
+ const [first, ...rest] = arrays;
+ return first.filter((value, index) => {
+ if (first.indexOf(value) !== index) return false;
+ return rest.every((values) => values.includes(value));
+ });
+}
+
+function minKnownNumber(values: Array<number | undefined>): number | undefined {
+ if (values.length === 0 || !values.every(isPositiveFiniteNumber)) return undefined;
+ return Math.min(...values);
+}
+
const VISION_MODEL_KEYWORDS = [
"gpt-4o",
"gpt-4.1",
@@ -320,6 +368,212 @@ export async function getUnifiedModelsResponse(
);
};
+ const getRegistryModel = (providerId: string, modelId: string) => {
+ const alias = providerIdToAlias[providerId] || PROVIDER_ID_TO_ALIAS[providerId] || providerId;
+ const providerModels = PROVIDER_MODELS[alias] || PROVIDER_MODELS[providerId] || [];
+ return providerModels.find((model) => model?.id === modelId) || null;
+ };
+
+ const getProviderPrefixes = (providerId: string, rawProvider: string) => {
+ const prefixes = new Set<string>([providerId, rawProvider, providerIdToAlias[providerId]]);
+ for (const [alias, mappedProviderId] of Object.entries(aliasToProviderId)) {
+ if (mappedProviderId === providerId) prefixes.add(alias);
+ }
+ return [...prefixes].filter(
+ (prefix): prefix is string => typeof prefix === "string" && prefix.length > 0
+ );
+ };
+
+ const getComboTargetModelId = (target: ComboCatalogTarget) => {
+ const rawProvider = typeof target.provider === "string" ? target.provider.trim() : "";
+ const modelStr = typeof target.modelStr === "string" ? target.modelStr.trim() : "";
+ if (!rawProvider || rawProvider === "unknown" || !modelStr) return null;
+
+ const providerId = resolveCanonicalProviderId(rawProvider);
+ if (!providerId || providerId === "unknown") return null;
+
+ for (const prefix of getProviderPrefixes(providerId, rawProvider)) {
+ const prefixWithSlash = `${prefix}/`;
+ if (modelStr.startsWith(prefixWithSlash)) {
+ const modelId = modelStr.slice(prefixWithSlash.length).trim();
+ return modelId ? { providerId, modelId } : null;
+ }
+ }
+
+ return { providerId, modelId: modelStr };
+ };
+
+ const getComboTargetCatalogMetadata = (
+ target: ComboCatalogTarget
+ ): ComboTargetCatalogMetadata | null => {
+ const targetModel = getComboTargetModelId(target);
+ if (!targetModel) return null;
+
+ const canonical = getCanonicalModelMetadata({
+ provider: targetModel.providerId,
+ model: targetModel.modelId,
+ });
+ if (!canonical) return null;
+
+ const source = canonical.metadata.source;
+ if (!source.providerRegistry && !source.staticSpec && !source.syncedCapability) return null;
+
+ const providerId = canonical.provider || targetModel.providerId;
+ const modelId = canonical.model || targetModel.modelId;
+ const synced = getSyncedCapability(providerId, modelId);
+ const spec = getModelSpec(modelId);
+ const registryModel = getRegistryModel(providerId, modelId);
+ const syncedInputModalities = parseJsonStringArray(synced?.modalities_input);
+ const syncedOutputModalities = parseJsonStringArray(synced?.modalities_output);
+
+ const syncedContext = isPositiveFiniteNumber(synced?.limit_context)
+ ? synced.limit_context
+ : undefined;
+ const registryContext = isPositiveFiniteNumber(registryModel?.contextLength)
+ ? registryModel.contextLength
+ : undefined;
+ const specContext = isPositiveFiniteNumber(spec?.contextWindow)
+ ? spec.contextWindow
+ : undefined;
+ const contextLength = syncedContext ?? registryContext ?? specContext;
+ const maxInputTokens = isPositiveFiniteNumber(synced?.limit_input)
+ ? synced.limit_input
+ : contextLength;
+ const maxOutputTokens = isPositiveFiniteNumber(synced?.limit_output)
+ ? synced.limit_output
+ : isPositiveFiniteNumber(spec?.maxOutputTokens)
+ ? spec.maxOutputTokens
+ : undefined;
+
+ const syncedVision =
+ typeof synced?.attachment === "boolean"
+ ? synced.attachment
+ : syncedInputModalities.length > 0 || syncedOutputModalities.length > 0
+ ? [...syncedInputModalities, ...syncedOutputModalities].some((entry) =>
+ entry.toLowerCase().includes("image")
+ )
+ : undefined;
+ const registryVision =
+ typeof registryModel?.supportsVision === "boolean"
+ ? registryModel.supportsVision
+ : undefined;
+ const specVision =
+ typeof spec?.supportsVision === "boolean" ? spec.supportsVision : undefined;
+ const knownVision = syncedVision ?? registryVision ?? specVision;
+
+ const inputModalities =
+ syncedInputModalities.length > 0
+ ? syncedInputModalities
+ : knownVision === true
+ ? ["text", "image"]
+ : undefined;
+ const outputModalities =
+ syncedOutputModalities.length > 0
+ ? syncedOutputModalities
+ : knownVision === true
+ ? ["text"]
+ : undefined;
+
+ const capabilities: Record<string, boolean> = {};
+ if (typeof synced?.tool_call === "boolean") {
+ capabilities.tool_calling = synced.tool_call;
+ } else if (typeof registryModel?.toolCalling === "boolean") {
+ capabilities.tool_calling = registryModel.toolCalling;
+ } else if (typeof spec?.supportsTools === "boolean") {
+ capabilities.tool_calling = spec.supportsTools;
+ }
+ if (typeof synced?.reasoning === "boolean") {
+ capabilities.reasoning = synced.reasoning;
+ } else if (typeof registryModel?.supportsReasoning === "boolean") {
+ capabilities.reasoning = registryModel.supportsReasoning;
+ } else if (typeof spec?.supportsThinking === "boolean") {
+ capabilities.reasoning = spec.supportsThinking;
+ }
+ if (typeof knownVision === "boolean") capabilities.vision = knownVision;
+ if (typeof synced?.attachment === "boolean") capabilities.attachment = synced.attachment;
+ if (typeof synced?.structured_output === "boolean") {
+ capabilities.structured_output = synced.structured_output;
+ }
+ if (typeof synced?.temperature === "boolean") capabilities.temperature = synced.temperature;
+ if (typeof synced?.reasoning === "boolean") {
+ capabilities.thinking = synced.reasoning;
+ } else if (typeof spec?.supportsThinking === "boolean") {
+ capabilities.thinking = spec.supportsThinking;
+ }
+
+ return {
+ ...(contextLength ? { contextLength } : {}),
+ ...(maxInputTokens ? { maxInputTokens } : {}),
+ ...(maxOutputTokens ? { maxOutputTokens } : {}),
+ ...(inputModalities && inputModalities.length > 0 ? { inputModalities } : {}),
+ ...(outputModalities && outputModalities.length > 0 ? { outputModalities } : {}),
+ capabilities,
+ };
+ };
+
+ const buildComboCatalogMetadata = (combo: Record<string, any>, allCombos: any[]) => {
+ const explicitContextLength = isPositiveFiniteNumber(combo.context_length)
+ ? combo.context_length
+ : undefined;
+
+ const baseMetadata = explicitContextLength ? { context_length: explicitContextLength } : {};
+ const targets = resolveNestedComboTargets(combo, allCombos) as ComboCatalogTarget[];
+ if (targets.length === 0) return baseMetadata;
+
+ const targetMetadata = targets.map((target) => getComboTargetCatalogMetadata(target));
+ if (targetMetadata.some((metadata) => metadata === null)) return baseMetadata;
+
+ const knownMetadata = targetMetadata as ComboTargetCatalogMetadata[];
+ const contextLength =
+ explicitContextLength ??
+ minKnownNumber(knownMetadata.map((metadata) => metadata.contextLength));
+ const maxInputTokens = minKnownNumber(
+ knownMetadata.map((metadata) => metadata.maxInputTokens)
+ );
+ const maxOutputTokens = minKnownNumber(
+ knownMetadata.map((metadata) => metadata.maxOutputTokens)
+ );
+
+ const inputModalities = knownMetadata.every(
+ (metadata) => Array.isArray(metadata.inputModalities) && metadata.inputModalities.length > 0
+ )
+ ? intersectStringArrays(knownMetadata.map((metadata) => metadata.inputModalities || []))
+ : [];
+ const outputModalities = knownMetadata.every(
+ (metadata) =>
+ Array.isArray(metadata.outputModalities) && metadata.outputModalities.length > 0
+ )
+ ? intersectStringArrays(knownMetadata.map((metadata) => metadata.outputModalities || []))
+ : [];
+
+ const capabilities: Record<string, boolean> = {};
+ for (const key of [
+ "tool_calling",
+ "reasoning",
+ "vision",
+ "attachment",
+ "structured_output",
+ "temperature",
+ "thinking",
+ ]) {
+ const values = knownMetadata.map((metadata) => metadata.capabilities[key]);
+ if (values.every((value): value is boolean => typeof value === "boolean")) {
+ const [first] = values;
+ if (values.every((value) => value === first)) capabilities[key] = first;
+ }
+ }
+
+ return {
+ ...baseMetadata,
+ ...(contextLength ? { context_length: contextLength } : {}),
+ ...(maxInputTokens ? { max_input_tokens: maxInputTokens } : {}),
+ ...(maxOutputTokens ? { max_output_tokens: maxOutputTokens } : {}),
+ ...(inputModalities.length > 0 ? { input_modalities: inputModalities } : {}),
+ ...(outputModalities.length > 0 ? { output_modalities: outputModalities } : {}),
+ ...(Object.keys(capabilities).length > 0 ? { capabilities } : {}),
+ };
+ };
+
// Collect models from active providers (or all if none active)
const models = [];
const timestamp = Math.floor(Date.now() / 1000);
@@ -327,30 +581,7 @@ export async function getUnifiedModelsResponse(
// Add combos first (they appear at the top) — only active ones
for (const combo of combos) {
if (combo.isActive === false || combo.isHidden === true) continue;
-
- // Calculate combo context length from its model targets.
- // OpenCode and other clients read context_length from the catalog; without it
- // they fall back to a conservative ~4000 token limit, causing truncation.
- const comboContextLength = Array.isArray(combo.models)
- ? combo.models
- .filter((step): step is ComboModelStep => step?.kind === "model" && Boolean(step.model))
- .map((step) => {
- const parsed = parseModel(step.model);
- const rawProvider = parsed.provider || step.providerId || "unknown";
- const provider = resolveCanonicalProviderId(rawProvider);
- const model = parsed.model || step.model;
- return getTokenLimit(provider, model);
- })
- .filter((limit): limit is number => typeof limit === "number" && limit > 0)
- .reduce((min, limit) => Math.min(min, limit), Infinity)
- : undefined;
-
- const effectiveContextLength =
- typeof combo.context_length === "number" && combo.context_length > 0
- ? combo.context_length
- : comboContextLength !== undefined && comboContextLength !== Infinity
- ? comboContextLength
- : undefined;
+ const comboMetadata = buildComboCatalogMetadata(combo, combos);
models.push({
id: combo.name,
@@ -360,7 +591,7 @@ export async function getUnifiedModelsResponse(
permission: [],
root: combo.name,
parent: null,
- ...(effectiveContextLength !== undefined ? { context_length: effectiveContextLength } : {}),
+ ...comboMetadata,
});
}
@@ -887,6 +1118,7 @@ export async function getUnifiedModelsResponse(
};
const enrichedModels = finalModels.map((model) => {
+ if (model.owned_by === "combo") return model;
const enriched = enrichCatalogModelEntry(model);
const fallbackContextLength = getDefaultContextFallback(enriched);
return fallbackContextLength
diff --git a/src/i18n/messages/bg.json b/src/i18n/messages/bg.json
index 615b046d2f..12d911c36c 100644
--- a/src/i18n/messages/bg.json
+++ b/src/i18n/messages/bg.json
@@ -4658,4 +4658,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/bn.json b/src/i18n/messages/bn.json
index f710c34764..cbb45ce673 100644
--- a/src/i18n/messages/bn.json
+++ b/src/i18n/messages/bn.json
@@ -4951,4 +4951,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/cs.json b/src/i18n/messages/cs.json
index 7017c186e9..ea7664394c 100644
--- a/src/i18n/messages/cs.json
+++ b/src/i18n/messages/cs.json
@@ -4658,4 +4658,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/da.json b/src/i18n/messages/da.json
index 5336c71495..1394d2f8c8 100644
--- a/src/i18n/messages/da.json
+++ b/src/i18n/messages/da.json
@@ -4658,4 +4658,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/de.json b/src/i18n/messages/de.json
index 0639d215b7..5e0a11e72e 100644
--- a/src/i18n/messages/de.json
+++ b/src/i18n/messages/de.json
@@ -4674,4 +4674,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/es.json b/src/i18n/messages/es.json
index 7ca0ac2d41..6acb1bf214 100644
--- a/src/i18n/messages/es.json
+++ b/src/i18n/messages/es.json
@@ -4658,4 +4658,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/fa.json b/src/i18n/messages/fa.json
index 7360b369ea..1aae9ab357 100644
--- a/src/i18n/messages/fa.json
+++ b/src/i18n/messages/fa.json
@@ -4951,4 +4951,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/fi.json b/src/i18n/messages/fi.json
index 4e75a03fbf..de451a629a 100644
--- a/src/i18n/messages/fi.json
+++ b/src/i18n/messages/fi.json
@@ -4658,4 +4658,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/fr.json b/src/i18n/messages/fr.json
index 44326e348d..61f8cf673e 100644
--- a/src/i18n/messages/fr.json
+++ b/src/i18n/messages/fr.json
@@ -4658,4 +4658,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/gu.json b/src/i18n/messages/gu.json
index 1204fe8a7f..6f6877fa45 100644
--- a/src/i18n/messages/gu.json
+++ b/src/i18n/messages/gu.json
@@ -4951,4 +4951,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/he.json b/src/i18n/messages/he.json
index e733ca551a..50fe996e44 100644
--- a/src/i18n/messages/he.json
+++ b/src/i18n/messages/he.json
@@ -4658,4 +4658,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/hi.json b/src/i18n/messages/hi.json
index 6213615420..a3030feb32 100644
--- a/src/i18n/messages/hi.json
+++ b/src/i18n/messages/hi.json
@@ -4658,4 +4658,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/hu.json b/src/i18n/messages/hu.json
index 0695fbfa9d..dde4b75d0a 100644
--- a/src/i18n/messages/hu.json
+++ b/src/i18n/messages/hu.json
@@ -4658,4 +4658,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/id.json b/src/i18n/messages/id.json
index 2b554d7928..78d6edc506 100644
--- a/src/i18n/messages/id.json
+++ b/src/i18n/messages/id.json
@@ -4662,4 +4662,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/in.json b/src/i18n/messages/in.json
index 001d49628d..93d1f02fc4 100644
--- a/src/i18n/messages/in.json
+++ b/src/i18n/messages/in.json
@@ -4951,4 +4951,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/it.json b/src/i18n/messages/it.json
index bdeb16216a..b13bbd59ef 100644
--- a/src/i18n/messages/it.json
+++ b/src/i18n/messages/it.json
@@ -4658,4 +4658,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/ja.json b/src/i18n/messages/ja.json
index 6f20f605ad..9245ca8f5b 100644
--- a/src/i18n/messages/ja.json
+++ b/src/i18n/messages/ja.json
@@ -4658,4 +4658,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/ko.json b/src/i18n/messages/ko.json
index b0200c0c6f..0dea5885de 100644
--- a/src/i18n/messages/ko.json
+++ b/src/i18n/messages/ko.json
@@ -4660,4 +4660,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/mr.json b/src/i18n/messages/mr.json
index a05d8fc744..1e5ce32f75 100644
--- a/src/i18n/messages/mr.json
+++ b/src/i18n/messages/mr.json
@@ -4951,4 +4951,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/ms.json b/src/i18n/messages/ms.json
index 2745bbf818..33cc008f9b 100644
--- a/src/i18n/messages/ms.json
+++ b/src/i18n/messages/ms.json
@@ -4658,4 +4658,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/nl.json b/src/i18n/messages/nl.json
index a72d112ff3..a9b77a8144 100644
--- a/src/i18n/messages/nl.json
+++ b/src/i18n/messages/nl.json
@@ -4658,4 +4658,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/no.json b/src/i18n/messages/no.json
index 5b5a755318..b86e81d080 100644
--- a/src/i18n/messages/no.json
+++ b/src/i18n/messages/no.json
@@ -4658,4 +4658,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/i18n/messages/phi.json b/src/i18n/messages/phi.json
index 35e1d22574..55f87b9e7a 100644
--- a/src/i18n/messages/phi.json
+++ b/src/i18n/messages/phi.json
@@ -4658,4 +4658,4 @@
"agentFeaturesContextLengthErrorInteger": "Context length must be a valid integer",
"agentFeaturesContextLengthErrorRange": "Context length must be between 1000 and 2000000"
}
-}
\ No newline at end of file
+}
diff --git a/src/lib/db/providers.ts b/src/lib/db/providers.ts
index f87953903a..9a023f40e2 100644
--- a/src/lib/db/providers.ts
+++ b/src/lib/db/providers.ts
@@ -442,9 +442,9 @@ export async function deleteProviderConnections(ids: string[]): Promise<number>
const deletedCount = db.transaction(() => {
const placeholders = ids.map(() => "?").join(",");
db.prepare(`DELETE FROM quota_snapshots WHERE connection_id IN (${placeholders})`).run(...ids);
- const result = db.prepare(
- `DELETE FROM provider_connections WHERE id IN (${placeholders})`
- ).run(...ids);
+ const result = db
+ .prepare(`DELETE FROM provider_connections WHERE id IN (${placeholders})`)
+ .run(...ids);
return result.changes ?? 0;
})();
diff --git a/src/shared/components/AutoRoutingBanner.tsx b/src/shared/components/AutoRoutingBanner.tsx
index 062de5c001..9aedb528ce 100644
--- a/src/shared/components/AutoRoutingBanner.tsx
+++ b/src/shared/components/AutoRoutingBanner.tsx
@@ -25,7 +25,7 @@ export default function AutoRoutingBanner() {
} catch {
// ignore localStorage errors (private mode, quotas)
}
-
+
setIsDismissed(true);
};
diff --git a/src/shared/utils/apiKeyPolicy.ts b/src/shared/utils/apiKeyPolicy.ts
index 5909d6b37f..c750f93361 100644
--- a/src/shared/utils/apiKeyPolicy.ts
+++ b/src/shared/utils/apiKeyPolicy.ts
@@ -17,9 +17,9 @@ import * as log from "@/sse/utils/logger";
import { checkRateLimit, RateLimitRule } from "./rateLimiter";
const DEFAULT_RATE_LIMITS: RateLimitRule[] = [
- { limit: 1000, window: 86400 }, // 1000 per day
- { limit: 5000, window: 604800 }, // 5000 per week
- { limit: 20000, window: 2592000 } // 20000 per month
+ { limit: 1000, window: 86400 }, // 1000 per day
+ { limit: 5000, window: 604800 }, // 5000 per week
+ { limit: 20000, window: 2592000 }, // 20000 per month
];
interface AccessSchedule {
@@ -187,7 +187,10 @@ export async function enforceApiKeyPolicy(
return {
apiKey,
apiKeyInfo,
- rejection: errorResponse(HTTP_STATUS.FORBIDDEN, "This API key is banned due to policy violations"),
+ rejection: errorResponse(
+ HTTP_STATUS.FORBIDDEN,
+ "This API key is banned due to policy violations"
+ ),
};
}
@@ -260,9 +263,10 @@ export async function enforceApiKeyPolicy(
// ── Check 5: Generic Multi-Window Rate Limits ──
if (apiKeyInfo.id) {
- const rulesToApply = (apiKeyInfo.rateLimits && apiKeyInfo.rateLimits.length > 0)
- ? [...apiKeyInfo.rateLimits]
- : [...DEFAULT_RATE_LIMITS];
+ const rulesToApply =
+ apiKeyInfo.rateLimits && apiKeyInfo.rateLimits.length > 0
+ ? [...apiKeyInfo.rateLimits]
+ : [...DEFAULT_RATE_LIMITS];
// Combine with legacy limits if they exist and custom rate limits aren't set
if (!apiKeyInfo.rateLimits || apiKeyInfo.rateLimits.length === 0) {
@@ -276,13 +280,16 @@ export async function enforceApiKeyPolicy(
const rateLimitResult = await checkRateLimit(apiKeyInfo.id, rulesToApply);
if (!rateLimitResult.allowed) {
- const failedWindowStr = rateLimitResult.failedWindow
- ? ` (${rateLimitResult.failedWindow}s window)`
+ const failedWindowStr = rateLimitResult.failedWindow
+ ? ` (${rateLimitResult.failedWindow}s window)`
: "";
return {
apiKey,
apiKeyInfo,
- rejection: errorResponse(HTTP_STATUS.RATE_LIMITED, `Request limit exceeded${failedWindowStr}. Please try again later.`),
+ rejection: errorResponse(
+ HTTP_STATUS.RATE_LIMITED,
+ `Request limit exceeded${failedWindowStr}. Please try again later.`
+ ),
};
}
}
diff --git a/src/shared/utils/rateLimiter.ts b/src/shared/utils/rateLimiter.ts
index 14fc5a89fe..e736c3b414 100644
--- a/src/shared/utils/rateLimiter.ts
+++ b/src/shared/utils/rateLimiter.ts
@@ -3,11 +3,10 @@ import Redis from "ioredis";
// Reuse existing REDIS_URL if set, or local redis via default docker-compose
// Use REDIS_URL from env (Docker/Production) or fallback to local redis
const REDIS_URL = process.env.REDIS_URL || "redis://localhost:6379";
-if (process.env.NODE_ENV === 'production' && !process.env.REDIS_URL) {
- console.warn('[REDIS] REDIS_URL is not set in production. Falling back to default.');
+if (process.env.NODE_ENV === "production" && !process.env.REDIS_URL) {
+ console.warn("[REDIS] REDIS_URL is not set in production. Falling back to default.");
}
-
let redisClient: Redis | null = null;
export function getRedisClient() {
@@ -17,9 +16,9 @@ export function getRedisClient() {
enableReadyCheck: false,
retryStrategy(times) {
return Math.min(times * 50, 2000); // Exponential backoff
- }
+ },
});
- redisClient.on('error', (err) => console.error('[REDIS] Error:', err.message));
+ redisClient.on("error", (err) => console.error("[REDIS] Error:", err.message));
}
return redisClient;
}
@@ -88,14 +87,17 @@ export function setRateLimiterTestMode(enabled: boolean) {
* Checks multi-window rate limits for an API key atomically via Redis.
*/
export async function checkRateLimit(
- keyId: string,
+ keyId: string,
rules: RateLimitRule[]
): Promise<RateLimitResult> {
if (!rules || rules.length === 0) return { allowed: true };
// ── In-memory mock for unit tests ──
- const isTestMode = explicitTestMode || process.env.NODE_ENV === "test" || process.env.DISABLE_SQLITE_AUTO_BACKUP === "true";
-
+ const isTestMode =
+ explicitTestMode ||
+ process.env.NODE_ENV === "test" ||
+ process.env.DISABLE_SQLITE_AUTO_BACKUP === "true";
+
if (isTestMode) {
const now = Math.floor(Date.now() / 1000);
for (const rule of rules) {
@@ -116,27 +118,25 @@ export async function checkRateLimit(
const redis = getRedisClient();
const args: (string | number)[] = [Math.floor(Date.now() / 1000)];
-
+
for (const rule of rules) {
args.push(rule.limit, rule.window);
}
try {
- const result = await redis.eval(
- RATE_LIMIT_SCRIPT,
- 1,
- `rl:api_key:${keyId}`,
- ...args
- ) as [number, number];
+ const result = (await redis.eval(RATE_LIMIT_SCRIPT, 1, `rl:api_key:${keyId}`, ...args)) as [
+ number,
+ number,
+ ];
if (result[0] === 0) {
return { allowed: false, failedWindow: result[1] };
}
-
+
return { allowed: true };
} catch (error) {
// Fail-open strategy if Redis goes down to prevent complete API outage
console.error("[RATE_LIMITER] Redis eval failed, bypassing rate limit:", error);
- return { allowed: true };
+ return { allowed: true };
}
}
diff --git a/tests/unit/api-key-regeneration.test.ts b/tests/unit/api-key-regeneration.test.ts
index d7624811ca..a492424498 100644
--- a/tests/unit/api-key-regeneration.test.ts
+++ b/tests/unit/api-key-regeneration.test.ts
@@ -41,15 +41,15 @@ test("regenerateApiKey creates a new key and invalidates the old one", async ()
const result = await apiKeysDb.regenerateApiKey(oldId);
assert.ok(result?.key);
const regenerated = result!.key;
-
+
assert.notEqual(regenerated, oldKey);
-
+
// New key should be valid
assert.equal(await apiKeysDb.validateApiKey(regenerated), true);
-
+
// Old key should be invalid
assert.equal(await apiKeysDb.validateApiKey(oldKey), false);
-
+
// Name and machineId should persist
const md = await apiKeysDb.getApiKeyMetadata(regenerated);
assert.equal(md?.name, "Regen Test");
diff --git a/tests/unit/base-executor-sanitize-effort.test.ts b/tests/unit/base-executor-sanitize-effort.test.ts
index af9b2f3b08..e07958f47c 100644
--- a/tests/unit/base-executor-sanitize-effort.test.ts
+++ b/tests/unit/base-executor-sanitize-effort.test.ts
@@ -1,9 +1,7 @@
import test from "node:test";
import assert from "node:assert/strict";
-const { sanitizeReasoningEffortForProvider } = await import(
- "../../open-sse/executors/base.ts"
-);
+const { sanitizeReasoningEffortForProvider } = await import("../../open-sse/executors/base.ts");
function makeLog() {
const messages: Array<[string, string]> = [];
@@ -20,12 +18,7 @@ test("sanitizeReasoningEffortForProvider: xiaomi-mimo downgrades xhigh → high"
reasoning_effort: "xhigh",
messages: [{ role: "user", content: "hi" }],
};
- const result = sanitizeReasoningEffortForProvider(
- body,
- "xiaomi-mimo",
- "mimo-v2.5-pro",
- log
- );
+ const result = sanitizeReasoningEffortForProvider(body, "xiaomi-mimo", "mimo-v2.5-pro", log);
assert.notEqual(result, body, "must return a new object when mutating");
assert.equal((result as any).reasoning_effort, "high");
assert.equal((result as any).model, "mimo-v2.5-pro", "other fields preserved");
diff --git a/tests/unit/models-catalog-route.test.ts b/tests/unit/models-catalog-route.test.ts
index 495b730abf..e0e33cdf52 100644
--- a/tests/unit/models-catalog-route.test.ts
+++ b/tests/unit/models-catalog-route.test.ts
@@ -24,10 +24,7 @@ async function resetStorage() {
fs.mkdirSync(TEST_DATA_DIR, { recursive: true });
}
-async function seedConnection(
- provider: string,
- overrides: Record<string, unknown> = {}
-) {
+async function seedConnection(provider: string, overrides: Record<string, unknown> = {}) {
return providersDb.createProviderConnection({
provider,
authType: (overrides.authType as string) || "apikey",
@@ -40,6 +37,29 @@ async function seedConnection(
});
}
+function capability(overrides = {}) {
+ return {
+ tool_call: null,
+ reasoning: null,
+ attachment: null,
+ structured_output: null,
+ temperature: null,
+ modalities_input: JSON.stringify([]),
+ modalities_output: JSON.stringify([]),
+ knowledge_cutoff: null,
+ release_date: null,
+ last_updated: null,
+ status: null,
+ family: null,
+ open_weights: null,
+ limit_context: null,
+ limit_input: null,
+ limit_output: null,
+ interleaved_field: null,
+ ...overrides,
+ };
+}
+
test.beforeEach(async () => {
await resetStorage();
});
@@ -231,6 +251,263 @@ test("v1 models catalog keeps only visible combos when no providers are active",
);
});
+test("v1 models catalog derives combo metadata from known targets conservatively", async () => {
+ try {
+ modelsDevSync.saveModelsDevCapabilities({
+ openai: {
+ "combo-alpha": capability({
+ tool_call: true,
+ reasoning: true,
+ attachment: true,
+ structured_output: true,
+ temperature: false,
+ modalities_input: JSON.stringify(["text", "image"]),
+ modalities_output: JSON.stringify(["text"]),
+ limit_context: 1000,
+ limit_input: 900,
+ limit_output: 120,
+ }),
+ },
+ gemini: {
+ "combo-beta": capability({
+ tool_call: true,
+ reasoning: true,
+ attachment: false,
+ structured_output: true,
+ temperature: false,
+ modalities_input: JSON.stringify(["text"]),
+ modalities_output: JSON.stringify(["text"]),
+ limit_context: 800,
+ limit_input: 700,
+ limit_output: 90,
+ }),
+ },
+ });
+
+ await combosDb.createCombo({
+ name: "metadata-router",
+ strategy: "priority",
+ models: ["openai/combo-alpha", "gemini/combo-beta"],
+ });
+
+ const response = await v1ModelsCatalog.getUnifiedModelsResponse(
+ new Request("http://localhost/api/v1/models")
+ );
+ const body = (await response.json()) as any;
+ const combo = body.data.find((item) => item.id === "metadata-router");
+
+ assert.equal(response.status, 200);
+ assert.ok(combo);
+ assert.equal(combo.context_length, 800);
+ assert.equal(combo.max_input_tokens, 700);
+ assert.equal(combo.max_output_tokens, 90);
+ assert.deepEqual(combo.input_modalities, ["text"]);
+ assert.deepEqual(combo.output_modalities, ["text"]);
+ assert.equal(combo.capabilities.structured_output, true);
+ assert.equal(combo.capabilities.temperature, false);
+ assert.equal(combo.capabilities.tool_calling, true);
+ assert.equal(combo.capabilities.reasoning, true);
+ assert.equal(combo.capabilities.thinking, true);
+ assert.equal("vision" in combo.capabilities, false);
+ assert.equal("attachment" in combo.capabilities, false);
+ assert.equal("architecture" in combo, false);
+ assert.equal("top_provider" in combo, false);
+ assert.equal("supported_parameters" in combo, false);
+ } finally {
+ modelsDevSync.saveModelsDevCapabilities({});
+ }
+});
+
+test("v1 models catalog lets explicit combo context override derived context", async () => {
+ try {
+ modelsDevSync.saveModelsDevCapabilities({
+ openai: {
+ "context-alpha": capability({
+ modalities_input: JSON.stringify(["text"]),
+ modalities_output: JSON.stringify(["text"]),
+ limit_context: 1000,
+ limit_input: 900,
+ limit_output: 120,
+ }),
+ },
+ gemini: {
+ "context-beta": capability({
+ modalities_input: JSON.stringify(["text"]),
+ modalities_output: JSON.stringify(["text"]),
+ limit_context: 800,
+ limit_input: 700,
+ limit_output: 90,
+ }),
+ },
+ });
+
+ const combo = await combosDb.createCombo({
+ name: "context-router",
+ strategy: "priority",
+ models: ["openai/context-alpha", "gemini/context-beta"],
+ });
+ await combosDb.updateCombo((combo as any).id, { context_length: 12345 });
+
+ const response = await v1ModelsCatalog.getUnifiedModelsResponse(
+ new Request("http://localhost/api/v1/models")
+ );
+ const body = (await response.json()) as any;
+ const listed = body.data.find((item) => item.id === "context-router");
+
+ assert.equal(response.status, 200);
+ assert.equal(listed.context_length, 12345);
+ assert.equal(listed.max_input_tokens, 700);
+ assert.equal(listed.max_output_tokens, 90);
+ } finally {
+ modelsDevSync.saveModelsDevCapabilities({});
+ }
+});
+
+test("v1 models catalog keeps unknown combo targets visible without guessed metadata", async () => {
+ await combosDb.createCombo({
+ name: "unknown-router",
+ strategy: "priority",
+ models: ["openai/no-known-metadata"],
+ });
+
+ const response = await v1ModelsCatalog.getUnifiedModelsResponse(
+ new Request("http://localhost/api/v1/models")
+ );
+ const body = (await response.json()) as any;
+ const combo = body.data.find((item) => item.id === "unknown-router");
+
+ assert.equal(response.status, 200);
+ assert.ok(combo);
+ assert.equal("context_length" in combo, false);
+ assert.equal("max_input_tokens" in combo, false);
+ assert.equal("max_output_tokens" in combo, false);
+ assert.equal("input_modalities" in combo, false);
+ assert.equal("output_modalities" in combo, false);
+ assert.equal("capabilities" in combo, false);
+});
+
+test("v1 models catalog aggregates nested combos and keeps hidden child combos unlisted", async () => {
+ try {
+ modelsDevSync.saveModelsDevCapabilities({
+ openai: {
+ "nested-alpha": capability({
+ modalities_input: JSON.stringify(["text"]),
+ modalities_output: JSON.stringify(["text"]),
+ limit_context: 1000,
+ limit_input: 900,
+ limit_output: 120,
+ }),
+ },
+ gemini: {
+ "nested-beta": capability({
+ modalities_input: JSON.stringify(["text"]),
+ modalities_output: JSON.stringify(["text"]),
+ limit_context: 800,
+ limit_input: 700,
+ limit_output: 90,
+ }),
+ },
+ });
+
+ await combosDb.createCombo({
+ name: "hidden-child-router",
+ strategy: "priority",
+ models: ["openai/nested-alpha", "gemini/nested-beta"],
+ isHidden: true,
+ });
+ await combosDb.createCombo({
+ name: "parent-router",
+ strategy: "priority",
+ models: ["hidden-child-router"],
+ });
+
+ const response = await v1ModelsCatalog.getUnifiedModelsResponse(
+ new Request("http://localhost/api/v1/models")
+ );
+ const body = (await response.json()) as any;
+ const parent = body.data.find((item) => item.id === "parent-router");
+
+ assert.equal(response.status, 200);
+ assert.ok(parent);
+ assert.equal(parent.context_length, 800);
+ assert.equal(parent.max_output_tokens, 90);
+ assert.equal(
+ body.data.some((item) => item.id === "hidden-child-router"),
+ false
+ );
+ } finally {
+ modelsDevSync.saveModelsDevCapabilities({});
+ }
+});
+
+test("v1 models catalog resolves provider aliases without corrupting slashful model ids", async () => {
+ try {
+ modelsDevSync.saveModelsDevCapabilities({
+ claude: {
+ "alias-model": capability({
+ modalities_input: JSON.stringify(["text"]),
+ modalities_output: JSON.stringify(["text"]),
+ limit_context: 2000,
+ limit_input: 1900,
+ limit_output: 200,
+ }),
+ },
+ openrouter: {
+ "Qwen/Qwen3-Coder": capability({
+ modalities_input: JSON.stringify(["text"]),
+ modalities_output: JSON.stringify(["text"]),
+ limit_context: 1600,
+ limit_input: 1500,
+ limit_output: 150,
+ }),
+ },
+ });
+
+ await combosDb.createCombo({
+ name: "alias-and-slash-router",
+ strategy: "priority",
+ models: [
+ { kind: "model", providerId: "claude", model: "cc/alias-model" },
+ { kind: "model", providerId: "openrouter", model: "Qwen/Qwen3-Coder" },
+ ],
+ });
+
+ const response = await v1ModelsCatalog.getUnifiedModelsResponse(
+ new Request("http://localhost/api/v1/models")
+ );
+ const body = (await response.json()) as any;
+ const combo = body.data.find((item) => item.id === "alias-and-slash-router");
+
+ assert.equal(response.status, 200);
+ assert.ok(combo);
+ assert.equal(combo.context_length, 1600);
+ assert.equal(combo.max_input_tokens, 1500);
+ assert.equal(combo.max_output_tokens, 150);
+ } finally {
+ modelsDevSync.saveModelsDevCapabilities({});
+ }
+});
+
+test("v1 models catalog does not final-enrich combo names as real models", async () => {
+ await combosDb.createCombo({
+ name: "gpt-5.5",
+ strategy: "priority",
+ models: ["openai/no-known-metadata"],
+ });
+
+ const response = await v1ModelsCatalog.getUnifiedModelsResponse(
+ new Request("http://localhost/api/v1/models")
+ );
+ const body = (await response.json()) as any;
+ const combo = body.data.find((item) => item.id === "gpt-5.5");
+
+ assert.equal(response.status, 200);
+ assert.ok(combo);
+ assert.equal(combo.owned_by, "combo");
+ assert.equal("max_output_tokens" in combo, false);
+ assert.equal("capabilities" in combo, false);
+});
+
test("v1 models catalog exposes claude alias and provider-prefixed built-in models with vision metadata", async () => {
await seedConnection("claude", {
authType: "oauth",
@@ -1016,7 +1293,6 @@ test("v1 models catalog falls back to getTokenLimit for models without registry
);
});
-
test("v1 models catalog prefers manual combo context_length over auto-calculated", async () => {
await seedConnection("openai", { name: "openai-manual-context" });
diff --git a/tests/unit/proxy-connection-test.test.ts b/tests/unit/proxy-connection-test.test.ts
index 358c715847..4f5c0f9a49 100644
--- a/tests/unit/proxy-connection-test.test.ts
+++ b/tests/unit/proxy-connection-test.test.ts
@@ -1,6 +1,9 @@
import test from "node:test";
import assert from "node:assert/strict";
-import { providerAllowsOptionalApiKey, SELF_HOSTED_CHAT_PROVIDER_IDS } from "@/shared/constants/providers";
+import {
+ providerAllowsOptionalApiKey,
+ SELF_HOSTED_CHAT_PROVIDER_IDS,
+} from "@/shared/constants/providers";
// ── Import test targets from connection test route ──────────────────────────
diff --git a/tests/unit/translator-claude-helper-thinking.test.ts b/tests/unit/translator-claude-helper-thinking.test.ts
index ca153a44d1..5c2f460f45 100644
--- a/tests/unit/translator-claude-helper-thinking.test.ts
+++ b/tests/unit/translator-claude-helper-thinking.test.ts
@@ -1,12 +1,9 @@
import test from "node:test";
import assert from "node:assert/strict";
-const { prepareClaudeRequest } = await import(
- "../../open-sse/translator/helpers/claudeHelper.ts"
-);
-const { DEFAULT_THINKING_CLAUDE_SIGNATURE } = await import(
- "../../open-sse/config/defaultThinkingSignature.ts"
-);
+const { prepareClaudeRequest } = await import("../../open-sse/translator/helpers/claudeHelper.ts");
+const { DEFAULT_THINKING_CLAUDE_SIGNATURE } =
+ await import("../../open-sse/config/defaultThinkingSignature.ts");
function multiTurnBodyWithoutThinkingBlock() {
return {
@@ -15,9 +12,7 @@ function multiTurnBodyWithoutThinkingBlock() {
{ role: "user", content: [{ type: "text", text: "hi" }] },
{
role: "assistant",
- content: [
- { type: "tool_use", id: "call_x", name: "ls", input: { path: "." } },
- ],
+ content: [{ type: "tool_use", id: "call_x", name: "ls", input: { path: "." } }],
},
{
role: "user",
@@ -72,9 +67,7 @@ test("prepareClaudeRequest: existing thinking block — redacted, signature repl
},
{
role: "user",
- content: [
- { type: "tool_result", tool_use_id: "call_y", content: "ok" },
- ],
+ content: [{ type: "tool_result", tool_use_id: "call_y", content: "ok" }],
},
],
};
@@ -94,15 +87,11 @@ test("prepareClaudeRequest: thinking disabled — no inject regardless of tool_u
{ role: "user", content: [{ type: "text", text: "hi" }] },
{
role: "assistant",
- content: [
- { type: "tool_use", id: "call_z", name: "ls", input: {} },
- ],
+ content: [{ type: "tool_use", id: "call_z", name: "ls", input: {} }],
},
{
role: "user",
- content: [
- { type: "tool_result", tool_use_id: "call_z", content: "ok" },
- ],
+ content: [{ type: "tool_result", tool_use_id: "call_z", content: "ok" }],
},
],
};
@@ -115,9 +104,7 @@ test("prepareClaudeRequest: thinking disabled — no inject regardless of tool_u
test("prepareClaudeRequest: thinking enabled + no tool_use — no inject (single-turn text)", () => {
const body = {
thinking: { type: "enabled", budget_tokens: 4096 },
- messages: [
- { role: "user", content: [{ type: "text", text: "hi" }] },
- ],
+ messages: [{ role: "user", content: [{ type: "text", text: "hi" }] }],
};
const result = prepareClaudeRequest(body as any, "kimi-coding");
const userContent = (result as any).messages[0].content;