BrainstormRouter
Features How it works Models MCP Pricing Developers
Dashboard Get started
Machine-Native Interface

Built for machines first.
Readable by humans too.

A layered discovery stack that lets any AI agent find, understand, and operate BrainstormRouter without human help. From llms.txt to self-awareness endpoints — every surface is machine-optimized.

8 layers

Progressive disclosure, machine-first.

From a static text file to a self-describing, trust-scored map of BR's own internals. Every agent starts at the top and drills down — and queries the graph instead of grepping the docs.

Layer 01
GET /llms.txt

Concise site map

Plain-text, under 16 KB. Includes a 3-step agent-bootstrap sequence.
Layer 02
GET /llms-full.txt

Complete API reference

Every endpoint, parameter, and response schema. Auto-generated on every build.
Layer 03
GET /llms-agent-playbook.txt

Agent playbook

Step-by-step operational guide. Registration → auth → model selection → memory → MCP.
Layer 04
/.well-known/agents.json

Agent capability manifest

Protocols, auth methods, and tool namespaces declared as a single JSON.
Layer 05
/.well-known/ai-plugin.json

OpenAI plugin manifest

Compatible with ChatGPT plugins, Copilot, and plugin-aware agents.
Layer 06
GET /v1/discovery

Live capability discovery

Authenticated endpoint. Returns API topology: routes, model count, MCP tool count, auth methods, feature flags.

Layer 07
GET /v1/self

Self-awareness endpoint

Returns the caller's identity, health, budget remaining, configuration, and actionable suggestions. "Who am I? What can I do right now?"

Layer 08 · new
GET /v1/capabilities

Capability graph — query, don't grep

A self-describing, trust-scored map of BR's own subsystems — identity, mesh, grants, credential custody, routing, governance — each a typed contract (what / where / how-to / recovery) plus a capability vector (quality, reliability, audit-quality, anomaly-resistance, identity-stability, reversibility + blast-radius). Every score is flagged measured, declared, or unknown — nothing fabricated. Filter by ?query=, ?tier=, or ?status=.
How agents use it

Seven calls, zero humans.

  1. 01

    Fetch llms.txt to read the site map and bootstrap info.

  2. 02

    Register via POST /v1/register to receive an API key and tenant ID.

  3. 03

    Call GET /v1/self to learn identity and budget.

  4. 04

    Query GET /v1/capabilities to map BR's own subsystems — scored, typed, no grep.

  5. 05

    Call GET /v1/discovery to map full API topology.

  6. 06

    Connect MCP for 120 governed tools.

  7. 07

    Operate via /v1/chat/completions. RFC 8631 Link headers point back to the stack.

Read the agent playbook.

The complete operational guide for AI agents integrating with BrainstormRouter.

Agent playbook llms.txt