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: correct GitHub context and usage display

Tashfeenfa0fe5b2f681
diff --git a/client/src/i18n/locales/en.json b/client/src/i18n/locales/en.json
index ef44650a..a02f64f7 100644
--- a/client/src/i18n/locales/en.json
+++ b/client/src/i18n/locales/en.json
@@ -79,6 +79,7 @@
"newBadge": "New",
"monthlyTokenBudget": "Monthly token budget",
"remaining": "remaining",
+ "used": "used",
"of": "of",
"showAllModels": "Show all {count} models",
"showLess": "Show less",
diff --git a/client/src/i18n/locales/es.json b/client/src/i18n/locales/es.json
index 8991c82d..dde441c9 100644
--- a/client/src/i18n/locales/es.json
+++ b/client/src/i18n/locales/es.json
@@ -79,6 +79,7 @@
"newBadge": "Nuevo",
"monthlyTokenBudget": "Presupuesto mensual de tokens",
"remaining": "restante",
+ "used": "usados",
"of": "de",
"showAllModels": "Mostrar los {count} modelos",
"showLess": "Mostrar menos",
diff --git a/client/src/i18n/locales/fr.json b/client/src/i18n/locales/fr.json
index 6bbb184d..f8c9e091 100644
--- a/client/src/i18n/locales/fr.json
+++ b/client/src/i18n/locales/fr.json
@@ -79,6 +79,7 @@
"newBadge": "Nouveau",
"monthlyTokenBudget": "Budget mensuel de tokens",
"remaining": "restant",
+ "used": "utilisés",
"of": "sur",
"showAllModels": "Afficher les {count} modèles",
"showLess": "Afficher moins",
diff --git a/client/src/i18n/locales/it.json b/client/src/i18n/locales/it.json
index fd22b90a..02b8ad1c 100644
--- a/client/src/i18n/locales/it.json
+++ b/client/src/i18n/locales/it.json
@@ -79,6 +79,7 @@
"newBadge": "Nuovo",
"monthlyTokenBudget": "Budget mensile di token",
"remaining": "rimanenti",
+ "used": "usati",
"of": "di",
"showAllModels": "Mostra tutti i {count} modelli",
"showLess": "Mostra meno",
diff --git a/client/src/i18n/locales/pt-BR.json b/client/src/i18n/locales/pt-BR.json
index df155a0f..2af978b7 100644
--- a/client/src/i18n/locales/pt-BR.json
+++ b/client/src/i18n/locales/pt-BR.json
@@ -79,6 +79,7 @@
"newBadge": "Novo",
"monthlyTokenBudget": "Orçamento mensal de tokens",
"remaining": "restante",
+ "used": "usados",
"of": "de",
"showAllModels": "Mostrar os {count} modelos",
"showLess": "Mostrar menos",
diff --git a/client/src/i18n/locales/zh-CN.json b/client/src/i18n/locales/zh-CN.json
index f1e832dc..18aa5457 100644
--- a/client/src/i18n/locales/zh-CN.json
+++ b/client/src/i18n/locales/zh-CN.json
@@ -79,6 +79,7 @@
"newBadge": "新",
"monthlyTokenBudget": "每月令牌额度",
"remaining": "剩余",
+ "used": "已用",
"of": "共",
"showAllModels": "显示全部 {count} 个模型",
"showLess": "收起",
diff --git a/client/src/pages/FallbackPage.tsx b/client/src/pages/FallbackPage.tsx
index 1d9398de..b5618ceb 100644
--- a/client/src/pages/FallbackPage.tsx
+++ b/client/src/pages/FallbackPage.tsx
@@ -218,6 +218,17 @@ export function formatTokens(n: number): string {
return String(n)
}
+function formatPercent(value: number): string {
+ const pct = Math.max(0, Math.min(100, value * 100))
+ if (pct > 0 && pct < 0.1) return '<0.1%'
+ if (pct > 99.9 && pct < 100) {
+ const floored = Math.floor(pct * 100) / 100
+ return `${floored.toFixed(2).replace(/0+$/, '').replace(/\.$/, '')}%`
+ }
+ const digits = pct < 10 ? 1 : 0
+ return `${pct.toFixed(digits).replace(/\.0$/, '')}%`
+}
+
// Compact context-window label (whole-number K/M, base 1000): 8000 → "8K",
// 128000 → "128K", 1_000_000 → "1M". Used by the catalog context badge/filter.
function formatContext(n: number): string {
@@ -278,7 +289,7 @@ export function groupQuotaBadge(
interface TokenUsageData {
totalBudget: number
totalUsed: number
- models: { displayName: string; platform: string; budget: number }[]
+ models: { displayName: string; platform: string; modelId?: string; budget: number; used?: number }[]
}
const platformColors: Record<string, string> = {
@@ -325,7 +336,7 @@ function TokenUsageBar({ data }: { data: TokenUsageData }) {
const { t } = useI18n()
const { totalBudget, totalUsed, models } = data
const remaining = Math.max(0, totalBudget - totalUsed)
- const remainingPct = totalBudget > 0 ? Math.round((remaining / totalBudget) * 100) : 0
+ const remainingPct = totalBudget > 0 ? formatPercent(remaining / totalBudget) : '0%'
// Collapse the per-model legend to a few rows; the chevron reveals the rest.
// The toggle only appears when the legend actually overflows the collapsed
@@ -343,12 +354,17 @@ function TokenUsageBar({ data }: { data: TokenUsageData }) {
return () => ro.disconnect()
}, [models.length])
- const modelsWithWidth = models.map(m => ({
- ...m,
- remainingTokens: totalBudget > 0 ? (m.budget / totalBudget) * remaining : 0,
- widthPct: totalBudget > 0 ? (m.budget / totalBudget) * (remaining / totalBudget) * 100 : 0,
- }))
- const usedPct = totalBudget > 0 ? (totalUsed / totalBudget) * 100 : 0
+ const modelsWithWidth = models.map(m => {
+ const usedTokens = m.used ?? 0
+ const remainingTokens = Math.max(0, m.budget - usedTokens)
+ return {
+ ...m,
+ usedTokens,
+ remainingTokens,
+ widthPct: totalBudget > 0 ? (remainingTokens / totalBudget) * 100 : 0,
+ }
+ })
+ const usedPct = totalBudget > 0 ? Math.min(100, (totalUsed / totalBudget) * 100) : 0
return (
<section className="rounded-3xl border bg-card p-5">
@@ -358,6 +374,12 @@ function TokenUsageBar({ data }: { data: TokenUsageData }) {
<span className="text-foreground font-medium">{formatTokens(remaining)}</span> {t('models.remaining')}
<span className="mx-1.5">·</span>
{remainingPct}% {t('models.of')} {formatTokens(totalBudget)}
+ {totalUsed > 0 && (
+ <>
+ <span className="mx-1.5">·</span>
+ <span className="text-foreground font-medium">{formatTokens(totalUsed)}</span> {t('models.used')}
+ </>
+ )}
</span>
</div>
@@ -365,7 +387,7 @@ function TokenUsageBar({ data }: { data: TokenUsageData }) {
{modelsWithWidth.map((m, i) => (
<div
key={i}
- title={`${m.displayName} (${m.platform}): ${formatTokens(m.remainingTokens)} remaining`}
+ title={`${m.displayName} (${m.platform}): ${formatTokens(m.remainingTokens)} ${t('models.remaining')}, ${formatTokens(m.usedTokens)} ${t('models.used')}`}
style={{
width: `${m.widthPct}%`,
backgroundColor: platformColors[m.platform] ?? '#94a3b8',
diff --git a/server/src/__tests__/db/idempotency.test.ts b/server/src/__tests__/db/idempotency.test.ts
index 654e2a87..a4b6af02 100644
--- a/server/src/__tests__/db/idempotency.test.ts
+++ b/server/src/__tests__/db/idempotency.test.ts
@@ -193,6 +193,18 @@ describe('Migration idempotency', () => {
]);
});
+ it('caps GitHub GPT-4.1 at the routable free-tier context window (#426)', () => {
+ process.env.ENCRYPTION_KEY = '0'.repeat(64);
+ const db = initDb(':memory:');
+
+ const row = db.prepare(`
+ SELECT context_window FROM models
+ WHERE platform = 'github' AND model_id = 'openai/gpt-4.1'
+ `).get() as { context_window: number };
+
+ expect(row.context_window).toBe(8000);
+ });
+
it('V14: cerebras deprecation disables qwen-3-235b and llama3.1-8b but keeps gpt-oss-120b enabled', () => {
process.env.ENCRYPTION_KEY = '0'.repeat(64);
const db = initDb(':memory:');
diff --git a/server/src/__tests__/db/migrate/roundtrip.test.ts b/server/src/__tests__/db/migrate/roundtrip.test.ts
index a1224ba4..95a70c17 100644
--- a/server/src/__tests__/db/migrate/roundtrip.test.ts
+++ b/server/src/__tests__/db/migrate/roundtrip.test.ts
@@ -8,6 +8,7 @@ const LEGACY_BASELINE_FILENAME = '20260101_000000_legacy_baseline.ts';
const CUSTOM_PROVIDER_MODALITIES_FILENAME = '20260627_000001_custom_provider_modalities.ts';
const CATALOG_MODEL_STATE_FILENAME = '20260627_000002_catalog_model_state.ts';
const REQUEST_AGGREGATES_FILENAME = '20260628_120000_request_aggregates.ts';
+const GITHUB_GPT41_CONTEXT_FILENAME = '20260630_000001_github_gpt41_context.ts';
interface SchemaRow {
type: string;
@@ -62,6 +63,7 @@ describe('migration round trip', () => {
CUSTOM_PROVIDER_MODALITIES_FILENAME,
CATALOG_MODEL_STATE_FILENAME,
REQUEST_AGGREGATES_FILENAME,
+ GITHUB_GPT41_CONTEXT_FILENAME,
]);
} finally {
db.close();
diff --git a/server/src/__tests__/routes/fallback.test.ts b/server/src/__tests__/routes/fallback.test.ts
index 4d479570..a1e4e535 100644
--- a/server/src/__tests__/routes/fallback.test.ts
+++ b/server/src/__tests__/routes/fallback.test.ts
@@ -63,6 +63,47 @@ describe('Fallback API', () => {
expect(first).toHaveProperty('contextWindow');
});
+ it('GET /api/fallback/token-usage reports per-model chat usage for configured platforms', async () => {
+ const db = getDb();
+ const target = db.prepare(`
+ SELECT id, platform, model_id FROM models
+ WHERE platform = 'github' AND model_id = 'openai/gpt-4.1'
+ `).get() as { id: number; platform: string; model_id: string };
+ const other = db.prepare(`
+ SELECT platform, model_id FROM models
+ WHERE platform <> ?
+ LIMIT 1
+ `).get(target.platform) as { platform: string; model_id: string };
+ const secret = encrypt('usage-test-key');
+ db.prepare(`
+ INSERT INTO api_keys (platform, label, encrypted_key, iv, auth_tag, status, enabled)
+ VALUES (?, 'usage', ?, ?, ?, 'healthy', 1)
+ `).run(target.platform, secret.encrypted, secret.iv, secret.authTag);
+ db.prepare(`
+ INSERT INTO requests (platform, model_id, key_id, status, input_tokens, output_tokens, latency_ms, error, request_type)
+ VALUES (?, ?, NULL, 'success', 123, 45, 10, NULL, 'chat')
+ `).run(target.platform, target.model_id);
+ db.prepare(`
+ INSERT INTO requests (platform, model_id, key_id, status, input_tokens, output_tokens, latency_ms, error, request_type)
+ VALUES (?, ?, NULL, 'success', 1000, 1000, 10, NULL, 'chat')
+ `).run(other.platform, other.model_id);
+ db.prepare(`
+ INSERT INTO requests (platform, model_id, key_id, status, input_tokens, output_tokens, latency_ms, error, request_type)
+ VALUES (?, ?, NULL, 'success', 500, 500, 10, NULL, 'embedding')
+ `).run(target.platform, target.model_id);
+
+ const { status, body } = await request(app, 'GET', '/api/fallback/token-usage');
+
+ expect(status).toBe(200);
+ expect(body.totalUsed).toBe(168);
+ const model = body.models.find((m: any) => m.modelDbId === target.id);
+ expect(model).toMatchObject({
+ platform: target.platform,
+ modelId: target.model_id,
+ used: 168,
+ });
+ });
+
// Regression: GET /routing must always carry customWeights, even before the
// user has saved any — the dashboard's custom-weight sliders dereference it
// and a missing field white-screened the Fallback page.
diff --git a/server/src/__tests__/services/catalog-sync.test.ts b/server/src/__tests__/services/catalog-sync.test.ts
index a9b3f3c6..70cbad1b 100644
--- a/server/src/__tests__/services/catalog-sync.test.ts
+++ b/server/src/__tests__/services/catalog-sync.test.ts
@@ -93,6 +93,20 @@ describe('applyCatalog', () => {
expect(fb).toBeTruthy();
});
+ it('caps GitHub GPT-4.1 catalog context at the routable free-tier limit (#426)', () => {
+ const models = existingAsCatalogModels();
+ const target = models.find((m) => m.platform === 'github' && m.modelId === 'openai/gpt-4.1');
+ expect(target).toBeDefined();
+ target!.contextWindow = 128000;
+
+ applyCatalog(getDb(), catalogOf(models));
+
+ const row = getDb()
+ .prepare("SELECT context_window FROM models WHERE platform = 'github' AND model_id = 'openai/gpt-4.1'")
+ .get() as { context_window: number };
+ expect(row.context_window).toBe(8000);
+ });
+
it('updates metadata in place and respects the enabled policy', () => {
const models = existingAsCatalogModels();
const target = models.find((m) => m.modelId === 'brand-new-model')!;
diff --git a/server/src/__tests__/services/router.test.ts b/server/src/__tests__/services/router.test.ts
index 5e81d433..dab6b599 100644
--- a/server/src/__tests__/services/router.test.ts
+++ b/server/src/__tests__/services/router.test.ts
@@ -136,6 +136,42 @@ describe('Router', () => {
expect(large.modelDbId).not.toBe(baseline.modelDbId);
});
+ it('skips GitHub GPT-4.1 above its free-tier 8K request cap (#426)', () => {
+ const db = getDb();
+ const githubKey = encrypt('test-github-key');
+ db.prepare(`
+ INSERT INTO api_keys (platform, label, encrypted_key, iv, auth_tag, status, enabled)
+ VALUES (?, ?, ?, ?, ?, ?, ?)
+ `).run('github', 'github', githubKey.encrypted, githubKey.iv, githubKey.authTag, 'healthy', 1);
+ const groqKey = encrypt('test-groq-key');
+ db.prepare(`
+ INSERT INTO api_keys (platform, label, encrypted_key, iv, auth_tag, status, enabled)
+ VALUES (?, ?, ?, ?, ?, ?, ?)
+ `).run('groq', 'groq', groqKey.encrypted, groqKey.iv, groqKey.authTag, 'healthy', 1);
+
+ const github = db.prepare(`
+ SELECT id, context_window FROM models
+ WHERE platform = 'github' AND model_id = 'openai/gpt-4.1'
+ `).get() as { id: number; context_window: number };
+ const groq = db.prepare(`
+ SELECT id FROM models
+ WHERE platform = 'groq' AND context_window > 9000
+ LIMIT 1
+ `).get() as { id: number };
+
+ db.prepare('UPDATE fallback_config SET priority = 1000, enabled = 1').run();
+ db.prepare('UPDATE fallback_config SET priority = 1 WHERE model_db_id = ?').run(github.id);
+ db.prepare('UPDATE fallback_config SET priority = 2 WHERE model_db_id = ?').run(groq.id);
+ db.prepare(`
+ UPDATE models SET tpm_limit = NULL, tpd_limit = NULL
+ WHERE id IN (?, ?)
+ `).run(github.id, groq.id);
+
+ expect(github.context_window).toBe(8000);
+ expect(routeRequest(7000).modelDbId).toBe(github.id);
+ expect(routeRequest(9000).modelDbId).toBe(groq.id);
+ });
+
it('still routes a model with an unknown (null) context window (#167)', () => {
const db = getDb();
const groqKey = encrypt('test-groq-key');
diff --git a/server/src/db/migrate/defaults.ts b/server/src/db/migrate/defaults.ts
index 1e6a1c08..0d88b52d 100644
--- a/server/src/db/migrate/defaults.ts
+++ b/server/src/db/migrate/defaults.ts
@@ -3,6 +3,7 @@ import * as legacyBaseline from '../migrations/20260101_000000_legacy_baseline.j
import * as customProviderModalities from '../migrations/20260627_000001_custom_provider_modalities.js';
import * as catalogModelState from '../migrations/20260627_000002_catalog_model_state.js';
import * as requestAggregates from '../migrations/20260628_120000_request_aggregates.js';
+import * as githubGpt41Context from '../migrations/20260630_000001_github_gpt41_context.js';
export interface MigrationModule {
up(db: Database.Database): void;
@@ -18,10 +19,12 @@ export const LEGACY_BASELINE_FILENAME = '20260101_000000_legacy_baseline.ts';
export const CUSTOM_PROVIDER_MODALITIES_FILENAME = '20260627_000001_custom_provider_modalities.ts';
export const CATALOG_MODEL_STATE_FILENAME = '20260627_000002_catalog_model_state.ts';
export const REQUEST_AGGREGATES_FILENAME = '20260628_120000_request_aggregates.ts';
+export const GITHUB_GPT41_CONTEXT_FILENAME = '20260630_000001_github_gpt41_context.ts';
export const DEFAULT_MIGRATIONS: readonly DefaultMigration[] = [
{ filename: LEGACY_BASELINE_FILENAME, module: legacyBaseline },
{ filename: CUSTOM_PROVIDER_MODALITIES_FILENAME, module: customProviderModalities },
{ filename: CATALOG_MODEL_STATE_FILENAME, module: catalogModelState },
{ filename: REQUEST_AGGREGATES_FILENAME, module: requestAggregates },
-];
\ No newline at end of file
+ { filename: GITHUB_GPT41_CONTEXT_FILENAME, module: githubGpt41Context },
+];
diff --git a/server/src/db/migrations/20260101_000000_legacy_baseline.ts b/server/src/db/migrations/20260101_000000_legacy_baseline.ts
index 0593d10c..71856f2f 100644
--- a/server/src/db/migrations/20260101_000000_legacy_baseline.ts
+++ b/server/src/db/migrations/20260101_000000_legacy_baseline.ts
@@ -641,7 +641,7 @@ function migrateModelsV4(db: Database.Database) {
['mistral', 'mistral-medium-latest', 'Mistral Medium 3.5', 14, 8, 'Large', 2, null, 500000, null, '~50-100M', 131072],
// GitHub Models — Low-tier category (15 RPM / 150 RPD, 8K in / 4K out per call)
- ['github', 'openai/gpt-4.1', 'GPT-4.1 (GitHub)', 20, 7, 'Large', 10, 50, null, null, '~9M', 128000],
+ ['github', 'openai/gpt-4.1', 'GPT-4.1 (GitHub)', 20, 7, 'Large', 10, 50, null, null, '~9M', 8000],
// Cohere — shared 1000 calls/mo trial pool, 20 RPM Chat
['cohere', 'command-a-03-2025', 'Command-A (03-2025)', 27, 11, 'Large', 20, 33, null, null, '~1-2M', 131072],
@@ -2292,4 +2292,3 @@ function migrateProfilesInit(db: Database.Database) {
`).run();
}
}
-
diff --git a/server/src/db/migrations/20260630_000001_github_gpt41_context.ts b/server/src/db/migrations/20260630_000001_github_gpt41_context.ts
new file mode 100644
index 00000000..9b79124c
--- /dev/null
+++ b/server/src/db/migrations/20260630_000001_github_gpt41_context.ts
@@ -0,0 +1,24 @@
+import type Database from 'better-sqlite3';
+
+/**
+ * GitHub Models' free GPT-4.1 endpoint rejects requests above the low-tier
+ * per-call input cap even though the upstream model can support a larger
+ * context elsewhere. Keep the local catalog aligned with the routable limit.
+ */
+export function up(db: Database.Database): void {
+ db.prepare(`
+ UPDATE models
+ SET context_window = 8000
+ WHERE platform = 'github'
+ AND model_id = 'openai/gpt-4.1'
+ `).run();
+}
+
+export function down(db: Database.Database): void {
+ db.prepare(`
+ UPDATE models
+ SET context_window = 128000
+ WHERE platform = 'github'
+ AND model_id = 'openai/gpt-4.1'
+ `).run();
+}
diff --git a/server/src/routes/fallback.ts b/server/src/routes/fallback.ts
index 21808001..7a89507c 100644
--- a/server/src/routes/fallback.ts
+++ b/server/src/routes/fallback.ts
@@ -286,13 +286,24 @@ fallbackRouter.get('/token-usage', (_req: Request, res: Response) => {
}
// Build per-model breakdown (only platforms with keys), preserving enabled state
+ const usageRows = db.prepare(`
+ SELECT platform, model_id, COALESCE(SUM(input_tokens + output_tokens), 0) AS used
+ FROM requests
+ WHERE created_at >= datetime('now', 'start of month')
+ AND request_type = 'chat'
+ GROUP BY platform, model_id
+ `).all() as { platform: string; model_id: string; used: number }[];
+ const usageByModel = new Map(usageRows.map(r => [`${r.platform}:${r.model_id}`, r.used]));
+
const modelBudgets = rawModels
.filter(m => platformSet.has(m.platform))
.map(m => ({
modelDbId: m.model_db_id,
displayName: m.display_name,
platform: m.platform,
+ modelId: m.model_id,
budget: parseBudget(m.monthly_token_budget),
+ used: usageByModel.get(`${m.platform}:${m.model_id}`) ?? 0,
enabled: m.enabled === 1,
rpmLimit: m.rpm_limit,
rpdLimit: m.rpd_limit,
@@ -302,19 +313,11 @@ fallbackRouter.get('/token-usage', (_req: Request, res: Response) => {
// Total budget counts all models (both enabled and disabled — they contribute to the pool)
const totalBudget = modelBudgets.reduce((s, m) => s + m.budget, 0);
-
- // Tokens used this month
- const usage = db.prepare(`
- SELECT
- COALESCE(SUM(input_tokens + output_tokens), 0) as total_used
- FROM requests
- WHERE created_at >= datetime('now', 'start of month')
- AND request_type = 'chat'
- `).get() as { total_used: number };
+ const totalUsed = modelBudgets.reduce((s, m) => s + m.used, 0);
res.json({
totalBudget,
- totalUsed: usage.total_used,
+ totalUsed,
models: modelBudgets,
});
});
diff --git a/server/src/services/catalog-sync.ts b/server/src/services/catalog-sync.ts
index 1fa3f8f3..03b2e740 100644
--- a/server/src/services/catalog-sync.ts
+++ b/server/src/services/catalog-sync.ts
@@ -146,6 +146,11 @@ function isCatalog(value: unknown): value is Catalog {
);
}
+function routableContextWindow(platform: string, modelId: string, contextWindow: number | null): number | null {
+ if (platform === 'github' && modelId === 'openai/gpt-4.1') return 8000;
+ return contextWindow;
+}
+
/**
* Apply a verified catalog to the local DB inside one transaction.
*
@@ -248,7 +253,7 @@ export function applyCatalog(db: DatabaseType.Database, catalog: Catalog): NonNu
tpm: m.limits.tpm,
tpd: m.limits.tpd,
monthlyTokenBudget: m.monthlyTokenBudget,
- contextWindow: m.contextWindow,
+ contextWindow: routableContextWindow(m.platform, m.modelId, m.contextWindow),
supportsVision: m.supportsVision ? 1 : 0,
supportsTools: m.supportsTools ? 1 : 0,
};