For AI & agents

Machine-readable files

Point your agent at these URLs. Content is generated from the same source and cached for 1h.

• llms.txtLLM discovery index (llmstxt.org). Link from your agent config or system prompt.
• skill.mdAgent skill file — workflows, APIs, packages, guardrails. Load in Cursor or Claude.
• /for-ai/llms.txtSame content as root llms.txt.
• /for-ai/skill.mdSame content as root skill.md.

On-chain (Base mainnet)

gridz.bio publishes to Base (chain 8453). Subjects remain *.gridz.eth; the resolver is read via Base RPC.

specs/deployments.md · base.easscan.org

Read a profile (start here)

curl -s "https://gridz.bio/api/profile/kevin.gridz.eth" | jq '.grid.cells[] | {key, value}'

Returns a signed gridz/1.0.0 JSON Grid. ok: false means nothing is published on-chain yet.

• Example API response
• Example rendered page

Verify attestations

import { verifyGrid } from "@gridz/core";

const res = await fetch("https://gridz.bio/api/profile/kevin.gridz.eth");
const { grid } = await res.json();
const report = await verifyGrid(grid);

See verification docs.

Publish (agents)

Do not call POST /api/publish from third-party agents. Sign locally, then publish via MCP, CLI, SDK, or 1Claw.

MCP config snippet

{
  "mcpServers": {
    "gridz": {
      "command": "npx",
      "args": ["-y", "@gridz/mcp"]
    }
  }
}

Agent identity cells

Documentation

• Find profileSearch by alias with live suggestions (on-chain + local drafts)
• API referenceGET /api/profile/{ens} — read any published profile
• Toolkit@gridz/core, mcp, sdk, sinks, oneclaw
• CLIgridz init, build, verify, publish
• ConceptsCells, attestations, sinks, verification
• VerificationHow to trust a Grid offline
• Using gridz.bioURLs, drafts, publish flow