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

fix(quota): share Requesty's 200 requests/day free cap across its models (#1328)

Thibault Jaigu950be09b8575
diff --git a/docs/en/architecture/02-quota-and-cooldown-engine.md b/docs/en/architecture/02-quota-and-cooldown-engine.md
index 0b60b9e0..a53b9d2e 100644
--- a/docs/en/architecture/02-quota-and-cooldown-engine.md
+++ b/docs/en/architecture/02-quota-and-cooldown-engine.md
@@ -45,7 +45,7 @@ Some providers enforce **one quota across the whole account**, not per model:
| Pool | Providers | Config |
|------|-----------|--------|
-| Daily requests | OpenRouter (1000/day free, 50/day <10 credits), ModelScope (2000→1800 margin) | `PROVIDER_DAILY_REQUEST_CAP_<PLATFORM>` |
+| Daily requests | OpenRouter (1000/day free, 50/day <10 credits), ModelScope (2000→1800 margin), Requesty (200→180 margin) | `PROVIDER_DAILY_REQUEST_CAP_<PLATFORM>` |
| Daily tokens | NavyAI (150K/day shared, model multipliers) | `PROVIDER_DAILY_TOKEN_CAP_<PLATFORM>` |
| Minute requests | NVIDIA NIM (40 RPM account-wide) | `PROVIDER_MINUTE_REQUEST_CAP_<PLATFORM>` |
diff --git a/docs/en/providers/01-supported-platforms.md b/docs/en/providers/01-supported-platforms.md
index 353765f4..d6c20f37 100644
--- a/docs/en/providers/01-supported-platforms.md
+++ b/docs/en/providers/01-supported-platforms.md
@@ -45,7 +45,7 @@ Of the 43 built-in platforms, **10 use dedicated adapters** and **33 ride `OpenA
| `bazaarlink` | BazaarLink | Keyed | OpenAI-compat | Only the `auto:free` route is cataloged — direct model ids are paid (#385). |
| `ainative` | AINative Studio | Keyed | OpenAI-compat | Advertises recurring ~10M tokens/month free allocation; quota treated as unverified until confirmed by a real account. |
| `aion` | Aion Labs | Keyed | OpenAI-compat | No-card free key; availability catalog-managed behind the 30-day age gate. |
-| `requesty` | Requesty | Keyed | OpenAI-compat | Router endpoint at `router.requesty.ai/v1`; free rows age into the monthly catalog. |
+| `requesty` | Requesty | Keyed | OpenAI-compat | Router endpoint at `router.requesty.ai/v1`; free rows age into the monthly catalog; the free tier is 200 requests/day shared across all free models. |
| `navy` | NavyAI | Keyed | OpenAI-compat | Free plan: 150K tokens/day and 20 RPM; live smoke tests require an explicit User-Agent header. |
| `nara` | NaraRouter | Keyed | OpenAI-compat | Free plan additionally requires Telegram channel/link verification; live-probed 2026-07-09. |
| `sealion` | SEA-LION (AI Singapore) | Keyed | OpenAI-compat | First-party API; Google sign-in, no card, no region wall; 10 RPM recurring free tier. |
diff --git a/docs/zh-cn/architecture/02-quota-and-cooldown-engine.md b/docs/zh-cn/architecture/02-quota-and-cooldown-engine.md
index 17d3bca4..72daae1c 100644
--- a/docs/zh-cn/architecture/02-quota-and-cooldown-engine.md
+++ b/docs/zh-cn/architecture/02-quota-and-cooldown-engine.md
@@ -45,7 +45,7 @@ releaseLease(leaseId) // finally 块里,幂等
| 池 | 提供方 | 配置 |
|------|-----------|--------|
-| 每日请求 | OpenRouter(免费 1000/天,<10 积分 50/天)、ModelScope(2000→1800 边际) | `PROVIDER_DAILY_REQUEST_CAP_<PLATFORM>` |
+| 每日请求 | OpenRouter(免费 1000/天,<10 积分 50/天)、ModelScope(2000→1800 边际)、Requesty(200→180 边际) | `PROVIDER_DAILY_REQUEST_CAP_<PLATFORM>` |
| 每日词元 | NavyAI(共享 150K/天,模型乘数) | `PROVIDER_DAILY_TOKEN_CAP_<PLATFORM>` |
| 每分钟请求 | NVIDIA NIM(账号级 40 RPM) | `PROVIDER_MINUTE_REQUEST_CAP_<PLATFORM>` |
diff --git a/docs/zh-cn/providers/01-supported-platforms.md b/docs/zh-cn/providers/01-supported-platforms.md
index c08270c6..c1647043 100644
--- a/docs/zh-cn/providers/01-supported-platforms.md
+++ b/docs/zh-cn/providers/01-supported-platforms.md
@@ -41,7 +41,7 @@
| `bazaarlink` | BazaarLink | 带密钥 | OpenAI 兼容 | 只有 `auto:free` 路由进了目录——直接指定模型 id 是付费的(#385)。 |
| `ainative` | AINative Studio | 带密钥 | OpenAI 兼容 | 宣称每月循环约 1000 万词元的免费配额;在真实账号确认之前按未核实处理。 |
| `aion` | Aion Labs | 带密钥 | OpenAI 兼容 | 无需信用卡的免费密钥;可用性在 30 天观察期之后交由目录管理。 |
-| `requesty` | Requesty | 带密钥 | OpenAI 兼容 | 路由端点位于 `router.requesty.ai/v1`;免费行随月度目录逐步纳入。 |
+| `requesty` | Requesty | 带密钥 | OpenAI 兼容 | 路由端点位于 `router.requesty.ai/v1`;免费行随月度目录逐步纳入;免费档为所有免费模型共享每天 200 次请求。 |
| `navy` | NavyAI | 带密钥 | OpenAI 兼容 | 免费计划:每日 15 万词元、20 RPM;线上冒烟测试需要显式 User-Agent 头。 |
| `nara` | NaraRouter | 带密钥 | OpenAI 兼容 | 免费计划额外要求 Telegram 频道/链接验证;2026 年 7 月 9 日做过线上探测。 |
| `sealion` | SEA-LION (AI Singapore) | 带密钥 | OpenAI 兼容 | 第一方 API;Google 登录、无需信用卡、没有地区墙;免费档每月循环 10 RPM。 |
diff --git a/server/src/__tests__/services/ratelimit.test.ts b/server/src/__tests__/services/ratelimit.test.ts
index 7ce2c9d7..dce4e8bb 100644
--- a/server/src/__tests__/services/ratelimit.test.ts
+++ b/server/src/__tests__/services/ratelimit.test.ts
@@ -282,6 +282,8 @@ describe('Rate Limiter', () => {
expect(getProviderDailyRequestCap('openrouter')).toBe(1000);
// ModelScope: 2000/day account-wide upstream, shipped as 1800 for margin (#581).
expect(getProviderDailyRequestCap('modelscope')).toBe(1800);
+ // Requesty: 200/day across all free models, shipped as 180 for margin.
+ expect(getProviderDailyRequestCap('requesty')).toBe(180);
expect(getProviderDailyRequestCap('groq')).toBeNull(); // no shared cap
process.env[ENV] = '50';
expect(getProviderDailyRequestCap('openrouter')).toBe(50);
@@ -306,6 +308,15 @@ describe('Rate Limiter', () => {
recordRequest('openrouter', 'model-c', testId);
expect(canUseProvider('openrouter', testId)).toBe(false); // 3 >= 3
});
+
+ it('shares the Requesty daily cap across its free models by default', () => {
+ for (let i = 0; i < 179; i++) {
+ recordRequest('requesty', i % 2 ? 'google/gemma-4-31b-it' : 'nvidia/nemotron-3-super-120b-a12b', testId);
+ }
+ expect(canUseProvider('requesty', testId)).toBe(true); // 179 < 180
+ recordRequest('requesty', 'mistral/leanstral-1-5', testId);
+ expect(canUseProvider('requesty', testId)).toBe(false); // 180 >= 180
+ });
});
describe('provider-wide daily token cap (NavyAI)', () => {
diff --git a/server/src/services/ratelimit.ts b/server/src/services/ratelimit.ts
index 0d5e2bab..fbc9c2be 100644
--- a/server/src/services/ratelimit.ts
+++ b/server/src/services/ratelimit.ts
@@ -564,6 +564,12 @@ const DEFAULT_PROVIDER_DAILY_REQUEST_CAPS: Record<string, number> = {
// models share it). 1800 leaves margin for validation probes and for drift
// between our ledger and the provider's own daily boundary (#581).
modelscope: 1800,
+ // Requesty's free tier is 200 requests/day per account, shared by every free
+ // model (requesty.ai/free-models, no card). The catalog rows each carry
+ // rpd 200, so without this gate the router could spend 200 per model. 180
+ // leaves the same margin as ModelScope for cooldown probes and drift from
+ // Requesty's own daily reset.
+ requesty: 180,
};
const DEFAULT_PROVIDER_DAILY_TOKEN_CAPS: Record<string, number> = {