feat(ai): add reranker touchpoint to LiteLLM proxy recipe (takeover of #2455 )

LiteLLM normalizes Cohere/Voyage/Jina rerank backends to the wire shape gateway.rerank() already speaks, so a reranker touchpoint on the litellm recipe makes any proxied rerank model reachable via `search.reranker.model litellm:<model>` with no adapter. Repairs from the original PR: - path is the LEAF '/rerank' (not '/v1/rerank'): LiteLLM serves both /rerank and /v1/rerank, and the recipe's setup_hint allows LITELLM_BASE_URL with or without the /v1 suffix — pinning '/v1/rerank' doubled to /v1/v1/rerank (404) on /v1-suffixed bases. - setup_hint appends the rerank guidance to master's current line instead of replacing it with a stale pre-/v1-suffix version. - cost_per_1m_tokens_usd stays undefined (pricing-unknown), matching the recipe's embedding/chat touchpoints and budget-tracker's deliberate litellm exclusion from the free-provider sets (a proxy can front a paid provider; the touchpoint field isn't consumed by rerank pricing anyway). Test drives gateway.rerank()'s real URL builder via the stubbed transport for both base-URL forms; the /v1-suffixed case fails with the original PR's path. Co-authored-by: ozp <ozp@users.noreply.github.com> Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-14 17:02:19 +00:00 · 2026-07-21 14:28:27 -07:00
13 changed files with 181 additions and 327 deletions
@@ -7,7 +7,7 @@
  <link rel="preconnect" href="https://fonts.googleapis.com" />
  <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin />
  <link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600&family=JetBrains+Mono:wght@400;500&display=swap" rel="stylesheet" />
-  <script type="module" crossorigin src="/admin/assets/index-CzLjRij_.js"></script>
+  <script type="module" crossorigin src="/admin/assets/index-CoGEje3-.js"></script>
  <link rel="stylesheet" crossorigin href="/admin/assets/index-GxkWX7v3.css">
 </head>
 <body>
@@ -39,7 +39,6 @@ export const api = {
  stats: () => apiFetch('/admin/api/stats'),
  health: () => apiFetch('/admin/api/health-indicators'),
  agents: () => apiFetch('/admin/api/agents'),
-  sources: () => apiFetch('/admin/api/sources'),
  requests: (page = 1, qs = '') => apiFetch(`/admin/api/requests?page=${page}${qs}`),
  apiKeys: () => apiFetch('/admin/api/api-keys'),
  createApiKey: (name: string) => apiFetch('/admin/api/api-keys', { method: 'POST', body: JSON.stringify({ name }) }),
@@ -23,16 +23,9 @@ interface Agent {
  total_requests: number;
  requests_today: number;
  token_ttl: number | null;
-  source_id: string | null;
-  federated_read: string[] | null;
  status: 'active' | 'revoked';
 }

-interface SourceOption {
-  id: string;
-  name: string | null;
-}
-
 interface ApiKey {
  id: string;
  name: string;
@@ -95,7 +88,6 @@ export function AgentsPage() {
                <th>Name</th>
                <th>Type</th>
                <th>Scopes</th>
-                <th>Source</th>
                <th>Status</th>
                <th>Requests</th>
                <th>Last Used</th>
@@ -116,9 +108,6 @@ export function AgentsPage() {
                      <span key={s} className={`badge badge-${s}`} style={{ marginRight: 4 }}>{s}</span>
                    ))}
                  </td>
-                  <td className="mono" style={{ fontSize: 12, color: 'var(--text-secondary)' }}>
-                    {a.auth_type === 'oauth' ? (a.source_id || 'default') : 'all'}
-                  </td>
                  <td>
                    <span className={`badge ${a.status === 'active' ? 'badge-success' : 'badge-danger'}`}>{a.status}</span>
                  </td>
@@ -268,29 +257,9 @@ function RegisterModal({ onClose, onRegistered }: {
    Object.fromEntries(ALLOWED_SCOPES_LIST.map(s => [s, s === 'read'])) as Record<Scope, boolean>,
  );
  const [ttl, setTtl] = useState('86400'); // 24h default
-  const [redirectUris, setRedirectUris] = useState('');
-  const [sources, setSources] = useState<SourceOption[]>([{ id: 'default', name: 'default' }]);
-  const [sourceId, setSourceId] = useState('default');
-  const [readSources, setReadSources] = useState<Record<string, boolean>>({ default: true });
  const [loading, setLoading] = useState(false);
  const [error, setError] = useState('');

-  useEffect(() => {
-    api.sources()
-      .then((rows: SourceOption[]) => {
-        const nextSources = rows.length > 0 ? rows : [{ id: 'default', name: 'default' }];
-        const nextSourceId = nextSources.some(s => s.id === sourceId) ? sourceId : nextSources[0].id;
-        setSources(nextSources);
-        setSourceId(nextSourceId);
-        setReadSources(prev => {
-          const next = Object.fromEntries(nextSources.map(s => [s.id, Boolean(prev[s.id])])) as Record<string, boolean>;
-          next[nextSourceId] = true;
-          return next;
-        });
-      })
-      .catch(() => {});
-  }, []);
-
  const ttlOptions = [
    { label: '1 hour', value: '3600' },
    { label: '24 hours', value: '86400' },
@@ -308,24 +277,11 @@ function RegisterModal({ onClose, onRegistered }: {
    try {
      // Use the CLI registration endpoint (POST to admin API)
      const selectedScopes = Object.entries(scopes).filter(([, v]) => v).map(([k]) => k).join(' ');
-      const federatedRead = sources.map(s => s.id).filter(id => readSources[id]);
-      if (federatedRead.length === 0) { setError('Select at least one read source'); setLoading(false); return; }
-      // #1036: redirect URIs, one per line. Mirrors the CLI convention
-      // (src/commands/auth.ts): presence of redirect URIs implies
-      // authorization_code + refresh_token grants unless the user had none.
-      const uris = redirectUris.split('\n').map(u => u.trim()).filter(Boolean);
      const res = await fetch('/admin/api/register-client', {
        method: 'POST',
        credentials: 'same-origin',
        headers: { 'Content-Type': 'application/json' },
-        body: JSON.stringify({
-          name: name.trim(),
-          scopes: selectedScopes,
-          tokenTtl: ttl === '0' ? 315360000 : Number(ttl),
-          source_id: sourceId,
-          federated_read: federatedRead,
-          ...(uris.length > 0 ? { redirectUris: uris, grantTypes: ['authorization_code', 'refresh_token'] } : {}),
-        }),
+        body: JSON.stringify({ name: name.trim(), scopes: selectedScopes, tokenTtl: ttl === '0' ? 315360000 : Number(ttl) }),
      });
      if (!res.ok) throw new Error('Registration failed');
      const data = await res.json();
@@ -356,42 +312,6 @@ function RegisterModal({ onClose, onRegistered }: {
            ))}
          </div>
        </div>
-        <div style={{ marginBottom: 16 }}>
-          <label>Write Source</label>
-          <select value={sourceId} onChange={e => {
-            const id = e.target.value;
-            setSourceId(id);
-            setReadSources(p => ({ ...p, [id]: true }));
-          }}
-            style={{ width: '100%', background: 'var(--bg-secondary)', color: 'var(--text-primary)', border: '1px solid var(--border)', borderRadius: 6, padding: '6px 10px', fontSize: 14 }}>
-            {sources.map(s => <option key={s.id} value={s.id}>{s.name && s.name !== s.id ? `${s.id} (${s.name})` : s.id}</option>)}
-          </select>
-        </div>
-        <div style={{ marginBottom: 16 }}>
-          <label>Read Sources</label>
-          <div className="checkbox-group">
-            {sources.map(s => (
-              <label key={s.id} className="checkbox-label">
-                <input
-                  type="checkbox"
-                  checked={Boolean(readSources[s.id])}
-                  onChange={e => setReadSources(p => ({ ...p, [s.id]: e.target.checked }))}
-                />
-                {s.name && s.name !== s.id ? `${s.id} (${s.name})` : s.id}
-              </label>
-            ))}
-          </div>
-        </div>
-        <div style={{ marginBottom: 16 }}>
-          <label>Redirect URIs <span style={{ color: 'var(--text-secondary)', fontWeight: 400 }}>(optional, one per line — enables authorization_code flow)</span></label>
-          <textarea
-            placeholder={'https://example.com/callback'}
-            value={redirectUris}
-            onChange={e => setRedirectUris(e.target.value)}
-            rows={2}
-            style={{ width: '100%', background: 'var(--bg-secondary)', color: 'var(--text-primary)', border: '1px solid var(--border)', borderRadius: 6, padding: '6px 10px', fontSize: 13, fontFamily: 'inherit', resize: 'vertical' }}
-          />
-        </div>
        <div style={{ marginBottom: 20 }}>
          <label>Token Lifetime</label>
          <select value={ttl} onChange={e => setTtl(e.target.value)}
@@ -608,8 +528,6 @@ function AgentDrawer({ agent, onClose, onRevoked }: { agent: Agent; onClose: ()
      client_name: agentName,
      auth_type: agent.auth_type,
      scope: agent.scope,
-      source_id: agent.source_id,
-      federated_read: agent.federated_read,
    }, null, 2),
  };

@@ -629,10 +547,6 @@ function AgentDrawer({ agent, onClose, onRevoked }: { agent: Agent; onClose: ()
          <span>{(agent.scope || '').split(' ').filter(Boolean).map(s => (
            <span key={s} className={`badge badge-${s}`} style={{ marginRight: 4 }}>{s}</span>
          ))}</span>
-          <span style={{ color: 'var(--text-secondary)' }}>Write Source</span>
-          <span className="mono">{agent.source_id || (isOAuth ? 'default' : 'all')}</span>
-          <span style={{ color: 'var(--text-secondary)' }}>Read Sources</span>
-          <span className="mono">{isOAuth ? (agent.federated_read && agent.federated_read.length > 0 ? agent.federated_read.join(', ') : (agent.source_id || 'default')) : 'all'}</span>
          <span style={{ color: 'var(--text-secondary)' }}>Registered</span>
          <span>{new Date(agent.created_at).toLocaleDateString()}</span>
          <span style={{ color: 'var(--text-secondary)' }}>Token TTL</span>
@@ -1,13 +1,13 @@
 // AUTO-GENERATED — do not edit by hand.
 // Run `bun run scripts/build-admin-embedded.ts` to regenerate.
-// Source: admin/dist/ at 2026-07-21.
+// Source: admin/dist/ at 2026-05-27.
 //
 // Bun resolves the file: imports to a path that works at runtime even
 // inside a compiled binary (`bun build --compile`). The manifest maps
 // the request path the express handler sees to (resolved-path, mime).

 // @ts-ignore — type: 'file' is Bun ESM, not in lib.d.ts
-import A_0_assets_index_CzLjRij__js from '../admin/dist/assets/index-CzLjRij_.js' with { type: 'file' };
+import A_0_assets_index_CoGEje3__js from '../admin/dist/assets/index-CoGEje3-.js' with { type: 'file' };
 // @ts-ignore — type: 'file' is Bun ESM, not in lib.d.ts
 import A_1_assets_index_GxkWX7v3_css from '../admin/dist/assets/index-GxkWX7v3.css' with { type: 'file' };
 // @ts-ignore — type: 'file' is Bun ESM, not in lib.d.ts
@@ -19,7 +19,7 @@ export interface AdminAsset {
 }

 export const ADMIN_ASSETS: Record<string, AdminAsset> = {
-  "/admin/assets/index-CzLjRij_.js": { path: A_0_assets_index_CzLjRij__js as unknown as string, mime: "application/javascript; charset=utf-8" },
+  "/admin/assets/index-CoGEje3-.js": { path: A_0_assets_index_CoGEje3__js as unknown as string, mime: "application/javascript; charset=utf-8" },
  "/admin/assets/index-GxkWX7v3.css": { path: A_1_assets_index_GxkWX7v3_css as unknown as string, mime: "text/css; charset=utf-8" },
  "/admin/index.html": { path: A_2_index_html as unknown as string, mime: "text/html; charset=utf-8" },
 };
@@ -39,7 +39,6 @@ import * as db from '../core/db.ts';
 import { sqlQueryForEngine, executeRawJsonb } from '../core/sql-query.ts';
 import { MinionQueue } from '../core/minions/queue.ts';
 import { isRetryableError } from '../core/retry-matcher.ts';
-import { normalizeAdminClientSourceScope } from '../core/admin-source-scope.ts';
 import {
  computeContentHash,
  validateIngestionEvent,
@@ -1087,7 +1086,6 @@ export async function runServeHttp(engine: BrainEngine, options: ServeHttpOption
      const oauthClients = await sql`
        SELECT c.client_id as id, c.client_name as name, 'oauth' as auth_type,
          c.grant_types, c.scope, c.created_at, c.token_ttl,
-          c.source_id, c.federated_read,
          CASE WHEN c.deleted_at IS NOT NULL THEN 'revoked' ELSE 'active' END as status,
          (SELECT max(created_at) FROM mcp_request_log WHERE token_name = c.client_id) as last_used_at,
          (SELECT count(*)::int FROM mcp_request_log WHERE token_name = c.client_id) as total_requests,
@@ -1097,7 +1095,6 @@ export async function runServeHttp(engine: BrainEngine, options: ServeHttpOption
      const legacyKeys = await sql`
        SELECT a.id, a.name, 'api_key' as auth_type,
          '{"bearer"}' as grant_types, 'read write admin' as scope, a.created_at, null as token_ttl,
-          null as source_id, null as federated_read,
          CASE WHEN a.revoked_at IS NOT NULL THEN 'revoked' ELSE 'active' END as status,
          a.last_used_at,
          (SELECT count(*)::int FROM mcp_request_log WHERE token_name = a.name) as total_requests,
@@ -1110,17 +1107,6 @@ export async function runServeHttp(engine: BrainEngine, options: ServeHttpOption
    }
  });

-  app.get('/admin/api/sources', requireAdmin, async (_req: Request, res: Response) => {
-    try {
-      const sources = await engine.listAllSources({ includeArchived: false });
-      const options = sources.map(s => ({ id: s.id, name: s.name }));
-      if (!options.some(s => s.id === 'default')) options.unshift({ id: 'default', name: 'default' });
-      res.json(options);
-    } catch (e) {
-      res.status(503).json({ error: 'service_unavailable' });
-    }
-  });
-
  // v0.38 Slice 4 — per-OAuth-client agent spend viewer. Pre-computes today's
  // spend (committed + pending reservations) per client so the Agents tab
  // can render a "$X / $Y today" cell. Read-side endpoint only — no mutation.
@@ -1454,8 +1440,6 @@ export async function runServeHttp(engine: BrainEngine, options: ServeHttpOption
      // missing, empty) and rejects the rest with a structured 400.
      const { name, tokenTtl, grantTypes, redirectUris, tokenEndpointAuthMethod } = req.body;
      const rawScopes = (req.body as Record<string, unknown>).scopes ?? (req.body as Record<string, unknown>).scope;
-      const rawSourceId = (req.body as Record<string, unknown>).source_id ?? (req.body as Record<string, unknown>).sourceId;
-      const rawFederatedRead = (req.body as Record<string, unknown>).federated_read ?? (req.body as Record<string, unknown>).federatedRead;
      if (!name) { res.status(400).json({ error: 'Name required' }); return; }
      let scopeString: string;
      try {
@@ -1486,23 +1470,8 @@ export async function runServeHttp(engine: BrainEngine, options: ServeHttpOption
        });
        return;
      }
-      let sourceScope: ReturnType<typeof normalizeAdminClientSourceScope>;
-      try {
-        const sources = await engine.listAllSources({ includeArchived: false });
-        sourceScope = normalizeAdminClientSourceScope(
-          rawSourceId,
-          rawFederatedRead,
-          sources.map(s => s.id),
-        );
-      } catch (e) {
-        res.status(400).json({
-          error: 'invalid_source_scope',
-          message: e instanceof Error ? e.message : String(e),
-        });
-        return;
-      }
      const result = await oauthProvider.registerClientManual(
-        name, grants, scopeString, uris, sourceScope.sourceId, sourceScope.federatedRead, validatedAuthMethod,
+        name, grants, scopeString, uris, 'default', undefined, validatedAuthMethod,
      );
      // Set per-client TTL if specified
      if (tokenTtl && Number(tokenTtl) > 0) {
@@ -1,51 +0,0 @@
-import { isValidSourceId } from './source-id.ts';
-
-export interface AdminClientSourceScope {
-  sourceId: string;
-  federatedRead: string[];
-}
-
-function parseFederatedReadInput(raw: unknown, sourceId: string): string[] {
-  if (raw === undefined || raw === null) return [sourceId];
-  if (typeof raw === 'string') {
-    const values = raw.split(',').map(s => s.trim()).filter(Boolean);
-    return values.length > 0 ? values : [sourceId];
-  }
-  if (Array.isArray(raw)) {
-    if (!raw.every(v => typeof v === 'string')) {
-      throw new Error('federated_read must be an array of source ids');
-    }
-    const values = raw.map(v => v.trim()).filter(Boolean);
-    return values.length > 0 ? values : [sourceId];
-  }
-  throw new Error('federated_read must be a string or array');
-}
-
-export function normalizeAdminClientSourceScope(
-  rawSourceId: unknown,
-  rawFederatedRead: unknown,
-  availableSourceIds: string[],
-): AdminClientSourceScope {
-  const knownSources = new Set(availableSourceIds);
-  const fallbackSourceId = knownSources.has('default') ? 'default' : (availableSourceIds[0] ?? 'default');
-  const sourceId = rawSourceId === undefined || rawSourceId === null || rawSourceId === ''
-    ? fallbackSourceId
-    : rawSourceId;
-  if (!isValidSourceId(sourceId)) {
-    throw new Error(`Invalid source_id: ${JSON.stringify(sourceId)}`);
-  }
-  if (knownSources.size > 0 && !knownSources.has(sourceId)) {
-    throw new Error(`Unknown source_id: ${sourceId}`);
-  }
-
-  const federatedRead = Array.from(new Set(parseFederatedReadInput(rawFederatedRead, sourceId)));
-  for (const id of federatedRead) {
-    if (!isValidSourceId(id)) {
-      throw new Error(`Invalid federated_read source_id: ${JSON.stringify(id)}`);
-    }
-    if (knownSources.size > 0 && !knownSources.has(id)) {
-      throw new Error(`Unknown federated_read source_id: ${id}`);
-    }
-  }
-  return { sourceId, federatedRead };
-}
@@ -56,6 +56,36 @@ export const litellmProxy: Recipe = {
      cost_per_1m_output_usd: undefined,
      price_last_verified: '2026-06-14',
    },
+    // LiteLLM normalizes Cohere / Voyage / Jina / etc. rerank backends to the
+    // same wire shape gbrain's gateway.rerank() already speaks (the
+    // ZeroEntropy/llama.cpp contract):
+    //   { model, query, documents, top_n } → { results: [{ index, relevance_score }] }
+    // So any rerank model the user registers in their LiteLLM config is
+    // reachable via `gbrain config set search.reranker.model litellm:<model>`
+    // with no request/response adapter — same as embeddings ride the proxy.
+    reranker: {
+      models: [], // user-provided; whatever rerank models the proxy serves
+      // No canonical default — the proxy defines its own model ids. The user
+      // sets search.reranker.model explicitly (mirrors the embedding
+      // touchpoint's user_provided_models contract).
+      default_model: '',
+      // The proxied backend bills (Cohere/Voyage/…); pricing-unknown is the
+      // honest state — same stance as this recipe's embedding/chat
+      // touchpoints and budget-tracker's deliberate litellm exclusion from
+      // the free-provider sets.
+      cost_per_1m_tokens_usd: undefined,
+      price_last_verified: '2026-06-27',
+      max_payload_bytes: 5_000_000,
+      // LEAF path only (matches llama-server-reranker's convention). LiteLLM
+      // serves both `/rerank` and `/v1/rerank`, and LITELLM_BASE_URL may be
+      // set with or without the `/v1` suffix (the setup_hint allows both), so
+      // the leaf form yields a valid route either way:
+      //   http://localhost:4000    + /rerank → /rerank        ✓
+      //   http://localhost:4000/v1 + /rerank → /v1/rerank     ✓
+      // Pinning '/v1/rerank' here would double to /v1/v1/rerank → 404 on
+      // /v1-suffixed bases.
+      path: '/rerank',
+    },
  },
-  setup_hint: 'Run LiteLLM (https://docs.litellm.ai) in front of any provider; set LITELLM_BASE_URL (include the /v1 suffix if your proxy serves the OpenAI route there, e.g. http://localhost:4000/v1) + pass --embedding-model litellm:<model> and --embedding-dimensions <N>.',
+  setup_hint: 'Run LiteLLM (https://docs.litellm.ai) in front of any provider; set LITELLM_BASE_URL (include the /v1 suffix if your proxy serves the OpenAI route there, e.g. http://localhost:4000/v1) + pass --embedding-model litellm:<model> and --embedding-dimensions <N>. For rerank: register a rerank model in LiteLLM and set search.reranker.model litellm:<model-name>.',
 };
@@ -1,44 +0,0 @@
-import { describe, expect, test } from 'bun:test';
-import { normalizeAdminClientSourceScope } from '../src/core/admin-source-scope.ts';
-
-describe('admin register-client source scope normalization', () => {
-  const sources = ['default', 'team-a', 'team-b'];
-
-  test('defaults write and read scope to default source', () => {
-    expect(normalizeAdminClientSourceScope(undefined, undefined, sources)).toEqual({
-      sourceId: 'default',
-      federatedRead: ['default'],
-    });
-  });
-
-  test('accepts snake-case body fields with multiple read sources', () => {
-    expect(normalizeAdminClientSourceScope('team-a', ['team-a', 'team-b'], sources)).toEqual({
-      sourceId: 'team-a',
-      federatedRead: ['team-a', 'team-b'],
-    });
-  });
-
-  test('accepts comma-separated federated_read for API callers', () => {
-    expect(normalizeAdminClientSourceScope('team-a', 'team-b, default,team-b', sources)).toEqual({
-      sourceId: 'team-a',
-      federatedRead: ['team-b', 'default'],
-    });
-  });
-
-  test('falls back to selected write source when federated_read is empty', () => {
-    expect(normalizeAdminClientSourceScope('team-b', [], sources)).toEqual({
-      sourceId: 'team-b',
-      federatedRead: ['team-b'],
-    });
-  });
-
-  test('rejects malformed source ids before registration', () => {
-    expect(() => normalizeAdminClientSourceScope('../secret', undefined, sources)).toThrow(/Invalid source_id/);
-    expect(() => normalizeAdminClientSourceScope('team-a', ['snake_case'], sources)).toThrow(/Invalid federated_read/);
-  });
-
-  test('rejects valid-looking but unregistered source ids', () => {
-    expect(() => normalizeAdminClientSourceScope('ghost', undefined, sources)).toThrow(/Unknown source_id/);
-    expect(() => normalizeAdminClientSourceScope('team-a', ['ghost'], sources)).toThrow(/Unknown federated_read/);
-  });
-});
@@ -0,0 +1,88 @@
+/**
+ * litellm-proxy reranker touchpoint smoke.
+ *
+ * Sibling of recipe-llama-server-reranker.test.ts. Pins the reranker
+ * touchpoint on the LiteLLM proxy recipe so:
+ *  - the touchpoint exists with the LEAF '/rerank' path (LiteLLM serves both
+ *    /rerank and /v1/rerank, so the leaf form is valid whether or not the
+ *    user's LITELLM_BASE_URL carries the /v1 suffix the setup_hint allows)
+ *  - a /v1-suffixed base URL does NOT produce /v1/v1/rerank (the original
+ *    community PR pinned '/v1/rerank' which 404s on /v1-suffixed bases)
+ *  - models: [] (user-provided; proxy defines the model ids)
+ *  - pricing stays undefined (proxy can front a paid provider — same honest
+ *    pricing-unknown stance as the embedding/chat touchpoints)
+ *
+ * The gateway.rerank() URL tests drive the real URL builder via the stubbed
+ * transport (same seam as test/ai/rerank.test.ts).
+ */
+
+import { describe, expect, test, afterEach } from 'bun:test';
+import { getRecipe } from '../../src/core/ai/recipes/index.ts';
+import {
+  configureGateway,
+  resetGateway,
+  rerank,
+  __setRerankTransportForTests,
+} from '../../src/core/ai/gateway.ts';
+
+afterEach(() => {
+  __setRerankTransportForTests(null);
+  resetGateway();
+});
+
+describe('recipe: litellm reranker touchpoint', () => {
+  test('declares reranker touchpoint with leaf /rerank path', () => {
+    const r = getRecipe('litellm')!;
+    const tp = r.touchpoints.reranker;
+    expect(tp).toBeDefined();
+    expect(tp!.path).toBe('/rerank');
+    expect(tp!.max_payload_bytes).toBe(5_000_000);
+  });
+
+  test('reranker touchpoint uses empty models[] for user-provided model ids', () => {
+    const r = getRecipe('litellm')!;
+    expect(r.touchpoints.reranker!.models).toEqual([]);
+  });
+
+  test('pricing stays undefined — proxy can front a paid provider', () => {
+    const r = getRecipe('litellm')!;
+    expect(r.touchpoints.reranker!.cost_per_1m_tokens_usd).toBeUndefined();
+  });
+
+  test('setup_hint keeps the /v1-suffix guidance AND mentions rerank', () => {
+    const r = getRecipe('litellm')!;
+    expect(r.setup_hint).toMatch(/\/v1 suffix/);
+    expect(r.setup_hint).toMatch(/search\.reranker\.model litellm:/);
+  });
+});
+
+describe('gateway.rerank() URL via litellm recipe', () => {
+  async function capturedRerankUrl(baseUrl?: string): Promise<string> {
+    configureGateway({
+      reranker_model: 'litellm:my-reranker',
+      env: {},
+      ...(baseUrl ? { base_urls: { litellm: baseUrl } } : {}),
+    });
+    let capturedUrl = '';
+    __setRerankTransportForTests(async (url) => {
+      capturedUrl = url;
+      return new Response(
+        JSON.stringify({ results: [{ index: 0, relevance_score: 0.9 }] }),
+        { status: 200, headers: { 'content-type': 'application/json' } },
+      );
+    });
+    await rerank({ query: 'q', documents: ['d'] });
+    return capturedUrl;
+  }
+
+  test('default base (no /v1 suffix) → /rerank', async () => {
+    const url = await capturedRerankUrl();
+    expect(url).toBe('http://localhost:4000/rerank');
+  });
+
+  test('/v1-suffixed base → /v1/rerank, NOT /v1/v1/rerank', async () => {
+    const url = await capturedRerankUrl('http://localhost:4000/v1');
+    expect(url).toBe('http://localhost:4000/v1/rerank');
+    expect(url).not.toContain('/v1/v1/');
+  });
+});
@@ -78,8 +78,6 @@ describe('v0.36.1.x #1077 — admin register-client supports PKCE public clients
    // (under either name) from req.body. Pin the fallback pattern so the
    // PKCE-fix regression contract stays load-bearing.
    expect(src).toMatch(/req\.body[^;]*scopes\s*\?\?\s*[^;]*scope\b/);
-    expect(src).toMatch(/req\.body[^;]*source_id\s*\?\?\s*[^;]*sourceId\b/);
-    expect(src).toMatch(/req\.body[^;]*federated_read\s*\?\?\s*[^;]*federatedRead\b/);
    // v0.41.3 (T4 atomicity fix, codex F4): admin endpoint now validates
    // tokenEndpointAuthMethod via the shared validator and passes it to
    // registerClientManual as a positional arg. Pre-v0.41.3 the route did
@@ -89,7 +87,6 @@ describe('v0.36.1.x #1077 — admin register-client supports PKCE public clients
    // UPDATE block (the regex deliberately asserts the post-insert UPDATE
    // is GONE).
    expect(src).toMatch(/validateTokenEndpointAuthMethod\(tokenEndpointAuthMethod\)/);
-    expect(src).toMatch(/normalizeAdminClientSourceScope\(/);
    expect(src).toMatch(/registerClientManual\([^)]*validatedAuthMethod[^)]*\)/);
    // Regression guard: post-insert UPDATE flipping client_secret_hash to
    // NULL based on a runtime check is exactly the non-atomic pattern T4
@@ -270,26 +267,3 @@ describe('v0.42.43.0 #2095 — volunteer-events sink + cycle purge wiring (struc
    expect(src).toMatch(/purged_volunteer_events_count/);
  });
 });
-
-describe('#1490 + #1036 — admin Register Agent form exposes source scope + redirect URIs', () => {
-  test('Agents.tsx RegisterModal sends source_id + federated_read in the POST body (#1490)', () => {
-    const src = readFileSync('admin/src/pages/Agents.tsx', 'utf8');
-    expect(src).toMatch(/source_id:\s*sourceId/);
-    expect(src).toMatch(/federated_read:\s*federatedRead/);
-    // The source list comes from the requireAdmin-gated endpoint, not a hardcoded 'default'.
-    expect(src).toMatch(/api\.sources\(\)/);
-  });
-
-  test('serve-http.ts serves /admin/api/sources behind requireAdmin (#1490)', () => {
-    const src = readFileSync('src/commands/serve-http.ts', 'utf8');
-    expect(src).toMatch(/app\.get\('\/admin\/api\/sources',\s*requireAdmin/);
-  });
-
-  test('Agents.tsx RegisterModal sends redirectUris + authorization_code grants when URIs are entered (#1036)', () => {
-    const src = readFileSync('admin/src/pages/Agents.tsx', 'utf8');
-    // One-per-line textarea split into an array...
-    expect(src).toMatch(/redirectUris\.split\('\\n'\)/);
-    // ...included in the body with the CLI's grant-type inference convention.
-    expect(src).toMatch(/redirectUris:\s*uris,\s*grantTypes:\s*\['authorization_code',\s*'refresh_token'\]/);
-  });
-});
@@ -125,30 +125,6 @@ describe('client registration', () => {
    expect(client!.client_name).toBe('test-agent');
  });

-  test('registerClientManual persists source_id and federated_read into AuthInfo', async () => {
-    await sql`
-      INSERT INTO sources (id, name) VALUES (${'team-a'}, ${'Team A'}), (${'team-b'}, ${'Team B'})
-      ON CONFLICT DO NOTHING
-    `;
-    const { clientId, clientSecret } = await provider.registerClientManual(
-      'source-scoped-test',
-      ['client_credentials'],
-      'read write',
-      [],
-      'team-a',
-      ['team-a', 'team-b'],
-    );
-
-    const tokens = await provider.exchangeClientCredentials(clientId, clientSecret!, 'read');
-    const authInfo = await provider.verifyAccessToken(tokens.access_token) as {
-      sourceId?: string;
-      allowedSources?: string[];
-    };
-
-    expect(authInfo.sourceId).toBe('team-a');
-    expect(authInfo.allowedSources).toEqual(['team-a', 'team-b']);
-  });
-
  test('getClient returns undefined for unknown client', async () => {
    const client = await provider.clientsStore.getClient('nonexistent');
    expect(client).toBeUndefined();