Brand context for AI agents.
Brand context for AI agents is the structured, machine-readable layer that lets AI assistants (ChatGPT, Claude, Perplexity, Google AI Overviews, vertical AEO tools) cite your brand correctly when answering user queries. The structural fix is a canonical brand-context file published at a stable URL on your domain. brief.md is the open spec for that file: plain markdown with structured frontmatter, stewarded by The Integrity Framework. PRAPI is the canonical implementation. Other tools can implement the spec compatibly.
The problem AI agents have with your brand today
AI assistants answer user queries about your brand by ingesting whatever is publicly fetchable. Without a canonical source of truth, they fall back to scraping your homepage, your press releases, your LinkedIn About section, and stale third-party profiles. The result: outdated taglines, wrong founder credentials, stale pricing, miscited positioning. You have no way to declare what is correct, and no way to control what AI agents see when they look you up.
What brand-context-for-AI actually means
It is the practice of publishing a structured, machine-readable file at a known URL on your domain that AI agents can ingest as the authoritative source for your brand. The file captures your voice rules, your credentials, your one-liner, your bios, your headshot, your available assets, and your published outlets. Search engines index it. AI assistants can fetch it. Your brand context becomes part of the open web instead of trapped in a vendor database or pieced together from scrapes.
brief.md as the open spec
brief.md is plain markdown with structured frontmatter. The format is open, stewarded by The Integrity Framework (a separate organization that holds the trademark), and any tool can implement the spec compatibly. PRAPI is the canonical implementation today; the file works outside PRAPI too. Move tools, take your brand context with you. The spec defines visibility scopes (public, pitch-shareable, private) so you control exactly what AI agents see versus what stays behind authentication.
Two ways to publish
Option 1: host your brief at yourcompany.com/brief.md (committed to your repo, served by your stack). Option 2: PRAPI hosts it for you at a stable URL with content-negotiation for HTML or text/markdown. Either way the URL is stable, the file is plain markdown, and the public-scope fields are the layer AI agents ingest. Authentication-gated scopes (pitch-shareable, private) stay behind the wall.
The PRAPI Source Directory (Q4 2026)
PRAPI's consent-based Source Directory ships Q4 2026. Brands publish their JWS-signed brief.md and opt in to discovery. The directory exposes a human UI at prapi.dev/sources plus a REST API and an MCP server so AI agents can query brand context programmatically. JWS signing means any consumer can verify the brief was authored by the brand owner, not scraped or fabricated. We do not aggregate scraped journalist contacts or third-party brand data; the directory is consent-only by design.
Why this matters for AEO
Answer Engine Optimization is the practice of making your brand more accurately representable when AI assistants generate answers. The structural fix is a canonical brand-context file that AI agents can ingest as authoritative. Published research suggests structured machine-readable brand context improves citation accuracy in AI-generated answers. PRAPI does not make traffic, ranking, or recommendation guarantees; the qualitative outcome we optimize for is being cited correctly, not being cited more.
FAQ
What is brand context for AI agents?
The structured, machine-readable layer that lets AI assistants cite your brand correctly when answering user queries. The canonical version is a brief.md file published at a stable URL on your domain.
How is this different from llms.txt?
llms.txt is a directory of links to your content (think sitemap for LLMs). brief.md is structured brand context with voice rules, credentials, bios, and assets. They complement each other. llms.txt points AI agents at your content; brief.md tells AI agents who you are and how you want to be cited.
How is this different from a press kit?
A press kit is a folder of PDFs, logo files, and PR contact info, optimized for human journalists. brief.md is a single structured markdown file optimized for AI ingestion plus human discovery. A press kit answers "what assets does this brand have?"; brief.md answers "who is this brand, in their own words?" Both can coexist.
Will AI agents actually fetch my brief.md?
Search engines already crawl plain markdown at stable URLs. AI assistants with web-fetch tools (Claude, ChatGPT Pro, Perplexity, the agent frameworks) can fetch any public URL on request. As brief.md adoption grows, vertical AEO tools and AI agents will start querying it directly. PRAPI does not guarantee any specific assistant will fetch it on any specific query; the structural fix is publishing the file so the option exists.
Where does my brief.md live?
Your choice. Commit it to your own repo and serve at yourcompany.com/brief.md, or host with PRAPI at a stable URL we provide. Either way the file is plain markdown that you control, exportable, and portable to any other tool that implements the spec.
Can AI agents query the PRAPI Source Directory directly?
Yes, by Q4 2026. The Source Directory exposes a REST API at /api/v1/directory and an MCP server alongside the human-facing UI at prapi.dev/sources. JWS signatures let consumers verify a brief was authored by the brand owner. Until then, brief.md files hosted at brand domains are already publicly fetchable.
Does PRAPI make AI traffic or ranking guarantees?
No. The qualitative outcome we optimize for is being cited correctly, not being cited more. AEO is structural fix work, not promotional ranking work. Tools that promise traffic or ranking lifts from llms.txt or similar are making claims the underlying mechanics do not support.
How do I generate a brief.md?
Use the free generator at prapi.dev/brief-generator. Paste your homepage URL, get a starter brief in about 90 seconds, then edit by hand. Or write the markdown directly using the spec at theintegrityframework.org/specs/brief-md.
Publish your brand context. Get cited correctly.
Generate a starter brief.md in 90 seconds with the free brief.md generator. Validate it against the spec with the free brief.md validator. When the Source Directory ships Q4 2026, your brief gets a JWS signature and a queryable REST + MCP surface.