Mission
The brand truth layer for the agentic web.
Encoded Brands gives organizations sovereignty over how agents understand, represent, recommend, sell, support, and transact with them.
We are not here to generate files. We are here to make brands legible.
Every agent in your stack is already making brand decisions. It decides what you are, what you sell, who you serve, and how you should sound. If you do not give it a canonical source, it rebuilds you from search results, stale press, scraped fragments, marketplace metadata, and model memory.
Encoded Brands exists so organizations can publish their own agent-readable source of truth. The Brand Context Protocol is the artifact. Sovereign brand understanding is the job.
Our mission is to make every serious brand accurately usable by autonomous agents, with ownership, provenance, and human approval built in.
The interface is changing. Brand control has to change with it.
For decades, brands were interpreted by people through websites, ads, search, social, packaging, retail, decks, and press. Now they are increasingly interpreted by agents. Agents compare vendors. They write recommendations. They summarize positioning. They answer customer questions. They will increasingly buy, book, negotiate, route, and renew.
That shift creates a simple problem. If agents become the interface, brand truth cannot live only in human-facing surfaces. It needs a standard address, a portable structure, and a way to separate owner-approved claims from inference.
- ExpressGive agents the words, facts, boundaries, proof, and context they need to represent the brand with fidelity.
- ProtectPrevent platforms, vendors, and models from reconstructing the brand from whatever data happens to be nearby.
- TransactLet commerce and marketing agents understand the brand before they act on its behalf.
Make the agentic web brand-safe, brand-owned, and commercially useful.
Agentic commerce will not work if every transaction begins with an agent guessing what a company means. Agentic marketing will not work if every campaign starts from a hallucinated brief. Agentic support will not work if every answer invents policy from fragments.
The most important thing we can build is not another interface. It is a durable layer of brand truth that any interface can read.
Manifesto
The brand should be read before the agent acts.
Agents are not waiting for permission to interpret brands. They already summarize, compare, recommend, rewrite, rank, and respond. They do this with whatever context they can find. Sometimes that context is good. Often it is residue.
A company should not be reconstructed from residue.
The web gave brands pages. Search gave brands snippets. Social gave brands feeds. Agents need something different: a canonical, owner-approved, agent-readable source of truth.
That source should be open. It should be portable. It should be small enough to fetch at runtime. It should separate fact from claim, claim from proof, proof from inference, and voice from vibe.
It should say what the brand is, what it sells, who it serves, what it believes, what it can claim, what it refuses, and how it wants to be represented.
It should belong to the brand, not the platform. It should be readable by every agent, not trapped inside one vendor. It should be boring in the way good infrastructure is boring: stable, inspectable, retrievable, and hard to misread.
/.well-known/brand.md /.well-known/brand/voice.md /.well-known/brand/values.md /.well-known/brand/boundaries.md /.well-known/brand/claims.md /.well-known/brand/representation.md /.well-known/brand/visual.md