Skip to content

7.4 billion tokens per month. 34 free LLM providers. 635 free model endpoints. All behind one /v1 endpoint, plus any custom OpenAI-compatible endpoint. Smart routing, automatic failover, encrypted keys. Personal experimentation only.

TypeScript

feat(providers): add xkiro.com as an OpenAI-compatible provider (#947) (#967)

tashdroidef4519a3d12c
diff --git a/client/src/components/keys/shared.tsx b/client/src/components/keys/shared.tsx
index d7376788..9ef39594 100644
--- a/client/src/components/keys/shared.tsx
+++ b/client/src/components/keys/shared.tsx
@@ -57,6 +57,7 @@ export const PLATFORMS: { value: Platform; label: string; url: string; keyless?:
{ value: 'sealion', label: 'SEA-LION (free key)', url: 'https://sea-lion.ai' },
{ value: 'orcarouter', label: 'OrcaRouter (free key)', url: 'https://www.orcarouter.ai' },
{ value: 'unorouter', label: 'UnoRouter (free key)', url: 'https://unorouter.com' },
+ { value: 'xkiro', label: 'xKiro (free key)', url: 'https://xkiro.com' },
// AnyAPI advertises 100K tokens/day free, but live testing on 2026-08-10
// could not get a single free-tier request served (see
// CATALOG-ANYAPI-SMOKE-2026-08-10 in the ops repo). No quota claim until
diff --git a/client/src/lib/routing.ts b/client/src/lib/routing.ts
index 0c6201be..95c96ca1 100644
--- a/client/src/lib/routing.ts
+++ b/client/src/lib/routing.ts
@@ -307,6 +307,7 @@ export const platformColors: Record<string, string> = {
sealion: '#0ea5e9',
orcarouter: '#f97316',
unorouter: '#ec4899',
+ xkiro: '#a855f7',
anyapi: '#0891b2',
modelscope: '#624aff',
aihorde: '#dc2626',
diff --git a/server/src/__tests__/lib/key-parser.test.ts b/server/src/__tests__/lib/key-parser.test.ts
index 2cf58fbc..db0f0a57 100644
--- a/server/src/__tests__/lib/key-parser.test.ts
+++ b/server/src/__tests__/lib/key-parser.test.ts
@@ -59,6 +59,7 @@ describe('key parser', () => {
expect(detectPlatform('ORCA_')).toBe('orcarouter');
expect(detectPlatform('UNOROUTER_')).toBe('unorouter');
expect(detectPlatform('UNO_ROUTER_')).toBe('unorouter');
+ expect(detectPlatform('XKIRO_')).toBe('xkiro');
expect(detectPlatform('MODELSCOPE_')).toBe('modelscope');
expect(detectPlatform('ANYAPI_')).toBe('anyapi');
expect(detectPlatform('ANY_API_')).toBe('anyapi');
diff --git a/server/src/__tests__/providers/openai-compat.test.ts b/server/src/__tests__/providers/openai-compat.test.ts
index 7abd044e..80df5915 100644
--- a/server/src/__tests__/providers/openai-compat.test.ts
+++ b/server/src/__tests__/providers/openai-compat.test.ts
@@ -519,6 +519,9 @@ describe('OpenAICompatProvider - platform instances', () => {
// unorouter's /v1/models requires auth (401 without a key), so default
// key validation works — no validateUrl override, unlike xkiro.
{ platform: 'unorouter', name: 'UnoRouter', baseUrl: 'https://api.unorouter.com/v1' },
+ // xkiro validates against /v1/usage (its /v1/models is public — 200 with no
+ // key), so it carries a validateUrl; chat routing is stock openai-compat.
+ { platform: 'xkiro', name: 'xKiro', baseUrl: 'https://api.xkiro.com/v1' },
// modelscope registers a ModelScopeProvider subclass (custom validateKey,
// see providers/modelscope.test.ts) but chat routing is stock openai-compat.
{ platform: 'modelscope', name: 'ModelScope', baseUrl: 'https://api-inference.modelscope.cn/v1' },
diff --git a/server/src/lib/key-parser.ts b/server/src/lib/key-parser.ts
index d7761ecb..44a73243 100644
--- a/server/src/lib/key-parser.ts
+++ b/server/src/lib/key-parser.ts
@@ -79,6 +79,7 @@ export const PREFIX_MAP: Record<string, string> = {
ORCA_: 'orcarouter',
UNOROUTER_: 'unorouter',
UNO_ROUTER_: 'unorouter',
+ XKIRO_: 'xkiro',
MODELSCOPE_: 'modelscope',
MODEL_SCOPE_: 'modelscope',
ANYAPI_: 'anyapi',
@@ -126,6 +127,7 @@ export const AUTH_JSON_PROVIDER_MAP: Record<string, string> = {
orca: 'orcarouter',
unorouter: 'unorouter',
'uno-router': 'unorouter',
+ xkiro: 'xkiro',
modelscope: 'modelscope',
'model-scope': 'modelscope',
anyapi: 'anyapi',
diff --git a/server/src/providers/index.ts b/server/src/providers/index.ts
index f12aebd6..6a6e4fbf 100644
--- a/server/src/providers/index.ts
+++ b/server/src/providers/index.ts
@@ -395,6 +395,27 @@ register(new OpenAICompatProvider({
baseUrl: 'https://api.unorouter.com/v1',
}));
+// xKiro (xkiro.com) — OpenAI-compatible gateway at api.xkiro.com/v1 (the
+// apex serves the same API today, but the docs name api.). Free key from
+// xkiro.com (no card). Live-probed 2026-08-23 with a free-plan key: /v1/usage
+// reports free_tokens limit_per_day=5,000,000; paid models answer an instant
+// 403 "premium model"/"requires a paid account", free ones (Mistral, MiniMax,
+// DeepSeek families) answer normally, Qwen was 503 upstream.
+// GET /v1/models answers 200 with NO key (public
+// catalog), so the default /v1/models key validation would be a false
+// positive — validateUrl points at /v1/usage instead, which 401s on a missing
+// or invalid ClientApiKey ("Invalid or disabled ClientApiKey"). Accepts
+// Authorization: Bearer or x-api-key. Catalog rows live in the hosted catalog
+// (premium now, free after the 30-day model-age gate); the ids in #947 are
+// unverified against a live account and are candidates for catalog authoring,
+// where a bad id is caught by the health check instead of shipped as a default.
+register(new OpenAICompatProvider({
+ platform: 'xkiro',
+ name: 'xKiro',
+ baseUrl: 'https://api.xkiro.com/v1',
+ validateUrl: 'https://api.xkiro.com/v1/usage',
+}));
+
// ModelScope (魔搭社区, Alibaba) — OpenAI-compatible inference API
// (api-inference.modelscope.cn/v1, Bearer auth). Free tier: 2000 requests/day
// account-wide. Token from modelscope.cn/my/myaccesstoken, BUT calls only work
diff --git a/server/src/routes/keys.ts b/server/src/routes/keys.ts
index bea1f066..78a025f9 100644
--- a/server/src/routes/keys.ts
+++ b/server/src/routes/keys.ts
@@ -32,7 +32,7 @@ const PLATFORMS = [
'google', 'groq', 'cerebras', 'bai', 'nvidia', 'mistral',
'openrouter', 'github', 'cohere', 'cloudflare', 'zhipu', 'ollama',
'kilo', 'pollinations', 'llm7', 'huggingface', 'opencode', 'ovh', 'agnes', 'reka', 'siliconflow',
- 'routeway', 'bazaarlink', 'ainative', 'aion', 'anyapi', 'requesty', 'navy', 'nara', 'sealion', 'orcarouter', 'unorouter', 'modelscope',
+ 'routeway', 'bazaarlink', 'ainative', 'aion', 'anyapi', 'requesty', 'navy', 'nara', 'sealion', 'orcarouter', 'unorouter', 'xkiro', 'modelscope',
'qianfan', 'volcengine', 'longcat', 'xfyun', 'aihorde', 'custom',
] as const;
diff --git a/server/src/services/provider-quota.ts b/server/src/services/provider-quota.ts
index 80670457..79c69bb6 100644
--- a/server/src/services/provider-quota.ts
+++ b/server/src/services/provider-quota.ts
@@ -150,6 +150,9 @@ function inferPoolForPlatform(platform: Platform, modelId?: string | null): stri
// 429 on every model for several minutes — so one pool, and a 429 on any
// model backs off the platform as a whole.
if (platform === 'unorouter') return 'unorouter::free';
+ // xkiro: one account-level allowance shared across its free models (the
+ // free tier is a per-account grant, not per-model), so one pool.
+ if (platform === 'xkiro') return 'xkiro::free';
// AnyAPI: the free tier is one 100K-tokens/day budget for the whole account,
// shared across every free/basic model — so one pool, not one per model.
if (platform === 'anyapi') return 'anyapi::free';
@@ -159,7 +162,7 @@ function inferPoolForPlatform(platform: Platform, modelId?: string | null): stri
}
function isSharedPool(platform: Platform): boolean {
- return ['openrouter', 'google', 'groq', 'cerebras', 'bai', 'sambanova', 'nvidia', 'mistral', 'github', 'cohere', 'cloudflare', 'zhipu', 'ollama', 'kilo', 'pollinations', 'llm7', 'huggingface', 'opencode', 'routeway', 'bazaarlink', 'ainative', 'aion', 'requesty', 'navy', 'nara', 'sealion', 'orcarouter', 'unorouter', 'anyapi', 'modelscope', 'aihorde'].includes(platform);
+ return ['openrouter', 'google', 'groq', 'cerebras', 'bai', 'sambanova', 'nvidia', 'mistral', 'github', 'cohere', 'cloudflare', 'zhipu', 'ollama', 'kilo', 'pollinations', 'llm7', 'huggingface', 'opencode', 'routeway', 'bazaarlink', 'ainative', 'aion', 'requesty', 'navy', 'nara', 'sealion', 'orcarouter', 'unorouter', 'xkiro', 'anyapi', 'modelscope', 'aihorde'].includes(platform);
}
type HeaderSpec = { metric: QuotaMetric; limit: string; remaining?: string; reset?: string; strategy?: QuotaResetStrategy };
diff --git a/shared/types.ts b/shared/types.ts
index c84a9a06..759805c0 100644
--- a/shared/types.ts
+++ b/shared/types.ts
@@ -141,6 +141,13 @@ export type Platform =
// validation works. Catalog rows live in the hosted catalog (premium now,
// free after the 30-day model-age gate).
| 'unorouter'
+ // xKiro (xkiro.com) — OpenAI-compatible gateway at api.xkiro.com/v1. Free key
+ // from xkiro.com (no card); free plan is 5M tokens/day on its free models,
+ // paid models 403 on a free key.
+ // /v1/models is public (200 with no key), so key validation must probe
+ // /v1/usage, which 401s on a missing/invalid ClientApiKey. Catalog rows live
+ // in the hosted catalog (premium now, free after the 30-day model-age gate).
+ | 'xkiro'
// ModelScope (魔搭社区, Alibaba) — OpenAI-compatible inference API
// (api-inference.modelscope.cn/v1). Free tier is 2000 requests/day
// account-wide, but calls only work after the ModelScope account is bound to