{"info":{"_postman_id":"6fd59f93-29c4-44d8-bd59-81309fd1f27d","name":"StatuteIQ Engine API (Preliminary)","description":"

$\"Brownsville$

\n\n

StatuteIQ Engine API

The Intelligent Compliance Engine, by Brownsville Advisory

\n
Navigate the complex world of compliance with confidence.
\n

Status: Preliminary. This documents the StatuteIQ engine's current and designed HTTP surface. It is not a GA contract; paths, schemas, and external authentication are still being finalized. Build-state is marked per request: LIVE (deployed today) versus DESIGNED (specified in the ENG-1 contract, not yet built).

What this is

The API service model is direct programmatic access to the StatuteIQ Compliance and Risk engine, custom-trained reasoning hosted on Google Vertex AI. It returns grounded, cited, schema-validated results, and it refuses to answer when evidence is insufficient (grounded:false) rather than fabricating one.

Quick start

Select an Environment (StatuteIQ - Local or StatuteIQ - Cloud Run) so baseUrl is set.
Set access_token. For local testing: gcloud auth print-identity-token. (Auth today is a Google OIDC identity-token bearer; external customer auth is TBD.)
Send a request, or run the Workflows folder to exercise the retrieve -> reason chain.

Folders

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n

Folder	Surface
Tools (LIVE)	The deployed tool surface: `/tools/list`, `/tools/invoke`, `/tools/classify`.
RAG API (DESIGNED)	`/api/retrieve` and `/api/reason` per ENG-1; live once the Knowledge Pool corpus loads.
Workflows (DESIGNED)	The canonical `retrieve -> reason` flow, with evidence passed between steps.

The grounding rule

When evidence is empty or insufficient, reasoning returns grounded:false with an empty result and empty citations. The engine never synthesizes clause text, award IDs, or docket numbers not present in the retrieved evidence. Always surface the returned citations and the pinned version to end users; that provenance is the product's core value.

Brownsville Advisory LLC · Washington, D.C. Metro Area · contact@brownsvilleadvisory.com

\n","schema":"https://schema.getpostman.com/json/collection/v2.0.0/collection.json","toc":[{"content":"StatuteIQ Engine API","slug":"statuteiq-engine-api"}],"owner":"56010874","collectionId":"6fd59f93-29c4-44d8-bd59-81309fd1f27d","publishedId":"2sBXwvHTsd","public":true,"customColor":{"top-bar":"FFFFFF","right-sidebar":"303030","highlight":"2E5FA3"},"publishDate":"2026-06-18T18:57:20.000Z"},"item":[{"name":"Tools (LIVE)","item":[{"name":"GET /tools/list","id":"af36c55e-68e6-4eab-855c-515e7327abb8","request":{"method":"GET","header":[{"key":"Content-Type","value":"application/json"},{"key":"Authorization","value":"Bearer "}],"url":"https://statuteiq-engine-REPLACE.a.run.app/tools/list","description":"

LIVE. List callable tools with input schemas.

LIVE. Invoke a named tool. Tools needing the RAG endpoints degrade to a cited corpus_pending result, never fabricated.

LIVE. Classify a natural-language query for routing (query intent, not document classification).

Currently deployed tool surface.

\n","_postman_id":"f503d592-6efb-4d0d-87d2-ff7e76ee9238","auth":{"type":"bearer","bearer":{"basicConfig":[{"key":"token","value":""}]},"isInherited":true,"source":{"_postman_id":"6fd59f93-29c4-44d8-bd59-81309fd1f27d","id":"6fd59f93-29c4-44d8-bd59-81309fd1f27d","name":"StatuteIQ Engine API (Preliminary)","type":"collection"}}},{"name":"RAG API (DESIGNED - ENG-1)","item":[{"name":"POST /api/retrieve","id":"2ae38468-af1b-408f-a104-1d8014dc8985","request":{"method":"POST","header":[{"key":"Content-Type","value":"application/json"},{"key":"Authorization","value":"Bearer "}],"body":{"mode":"raw","raw":"{\n \"query\": \"GAO sustain rate for organizational conflict of interest protests\",\n \"tags\": [\n \"protest\",\n \"GAO\",\n \"COFC\"\n ],\n \"limit\": 30,\n \"profile\": \"win-loss\",\n \"classification\": \"public\"\n}","options":{"raw":{"language":"json"}}},"url":"https://statuteiq-engine-REPLACE.a.run.app/api/retrieve","description":"

DESIGNED. Hybrid retrieval (vector + BM25 -> cross-encoder rerank) over the Knowledge Pool. classification enforces the no-train boundary.

\n","auth":{"type":"bearer","bearer":{"basicConfig":[{"key":"token","value":""}]},"isInherited":true,"source":{"_postman_id":"6fd59f93-29c4-44d8-bd59-81309fd1f27d","id":"6fd59f93-29c4-44d8-bd59-81309fd1f27d","name":"StatuteIQ Engine API (Preliminary)","type":"collection"}},"urlObject":{"path":["api","retrieve"],"host":["https://statuteiq-engine-REPLACE.a.run.app"],"query":[],"variable":[]}},"response":[],"_postman_id":"2ae38468-af1b-408f-a104-1d8014dc8985"},{"name":"POST /api/reason","id":"a584a340-06fc-4dd8-83f9-8648d7fc200c","request":{"method":"POST","header":[{"key":"Content-Type","value":"application/json"},{"key":"Authorization","value":"Bearer "}],"body":{"mode":"raw","raw":"{\n \"prompt\": \"Identify winning-language patterns for OCI mitigation.\",\n \"systemInstruction\": \"You are a federal capture analyst. Cite every pattern to source.\",\n \"evidence\": {\n \"chunks\": [],\n \"retrieval\": {\n \"profile\": \"win-loss\",\n \"returned\": 0,\n \"tagsMatched\": []\n }\n },\n \"responseShape\": \"winning-language\"\n}","options":{"raw":{"language":"json"}}},"url":"https://statuteiq-engine-REPLACE.a.run.app/api/reason","description":"

DESIGNED. Vertex-grounded reasoning over evidence, validated against a named responseShape. Returns the validated object + citations + grounded flag.

Specified in ENG-1; not yet built. Lights up once the Knowledge Pool corpus is loaded.

\n","_postman_id":"8240f060-6a88-447b-a5a8-b4b2674db858","auth":{"type":"bearer","bearer":{"basicConfig":[{"key":"token","value":""}]},"isInherited":true,"source":{"_postman_id":"6fd59f93-29c4-44d8-bd59-81309fd1f27d","id":"6fd59f93-29c4-44d8-bd59-81309fd1f27d","name":"StatuteIQ Engine API (Preliminary)","type":"collection"}}},{"name":"Workflows (DESIGNED)","item":[{"name":"1 - Retrieve evidence","event":[{"listen":"test","script":{"type":"text/javascript","exec":["// Flow step 1: stash retrieved evidence for the Reason step.","if (pm.response.code === 200) {"," try { pm.collectionVariables.set('evidence', JSON.stringify(pm.response.json())); }"," catch (e) { pm.collectionVariables.set('evidence', '{\"chunks\":[],\"retrieval\":{}}'); }","}"]}}],"id":"afc79097-7b79-4d5f-bc9b-fd4b02505742","request":{"method":"POST","header":[{"key":"Content-Type","value":"application/json"},{"key":"Authorization","value":"Bearer "}],"body":{"mode":"raw","raw":"{\n \"query\": \"GAO sustain rate for organizational conflict of interest protests\",\n \"tags\": [\n \"protest\",\n \"GAO\",\n \"COFC\"\n ],\n \"limit\": 30,\n \"profile\": \"win-loss\",\n \"classification\": \"public\"\n}","options":{"raw":{"language":"json"}}},"url":"https://statuteiq-engine-REPLACE.a.run.app/api/retrieve","description":"

Flow step 1. On 200, saves the response to {\"chunks\":[],\"retrieval\":{}} for step 2.

\n","auth":{"type":"bearer","bearer":{"basicConfig":[{"key":"token","value":""}]},"isInherited":true,"source":{"_postman_id":"6fd59f93-29c4-44d8-bd59-81309fd1f27d","id":"6fd59f93-29c4-44d8-bd59-81309fd1f27d","name":"StatuteIQ Engine API (Preliminary)","type":"collection"}},"urlObject":{"path":["api","retrieve"],"host":["https://statuteiq-engine-REPLACE.a.run.app"],"query":[],"variable":[]}},"response":[],"_postman_id":"afc79097-7b79-4d5f-bc9b-fd4b02505742"},{"name":"2 - Reason over evidence","event":[{"listen":"test","script":{"type":"text/javascript","exec":["// Flow step 2: sanity-check the grounded contract.","try { pm.test('response has grounded flag', () => pm.expect(pm.response.json()).to.have.property('grounded')); } catch(e){}"]}}],"id":"1930da59-338d-4819-9194-fdf4e4bc9134","request":{"method":"POST","header":[{"key":"Content-Type","value":"application/json"},{"key":"Authorization","value":"Bearer "}],"body":{"mode":"raw","raw":"{\n \"prompt\": \"Identify winning-language patterns for OCI mitigation, grounded only in the retrieved evidence.\",\n \"systemInstruction\": \"You are a federal capture analyst. Cite every pattern to source; if evidence is insufficient, return grounded:false.\",\n \"evidence\": {\"chunks\":[],\"retrieval\":{}},\n \"responseShape\": \"winning-language\"\n}","options":{"raw":{"language":"json"}}},"url":"https://statuteiq-engine-REPLACE.a.run.app/api/reason","description":"

Flow step 2. Consumes {\"chunks\":[],\"retrieval\":{}} from step 1 (passes the retrieved chunks straight through).

The canonical engine flow per ENG-1: retrieve evidence, then reason over it. Run this folder with the Collection Runner. Step 1 stashes its response into the {\"chunks\":[],\"retrieval\":{}} variable; Step 2 passes it straight into /api/reason. Both endpoints are DESIGNED (not yet live).

\n","_postman_id":"8f4ab217-b9cf-43fc-831b-bdf24d31c806","auth":{"type":"bearer","bearer":{"basicConfig":[{"key":"token","value":""}]},"isInherited":true,"source":{"_postman_id":"6fd59f93-29c4-44d8-bd59-81309fd1f27d","id":"6fd59f93-29c4-44d8-bd59-81309fd1f27d","name":"StatuteIQ Engine API (Preliminary)","type":"collection"}}}],"auth":{"type":"bearer","bearer":{"basicConfig":[{"key":"token","value":""}]}},"variable":[{"key":"baseUrl","value":"https://statuteiq-engine-REPLACE.a.run.app","type":"string"},{"key":"access_token","value":"","type":"string"},{"key":"evidence","value":"{\"chunks\":[],\"retrieval\":{}}","type":"string"}]}