Build on the brain.

curl https://mcp.falkster.ai/api/v1/ask \
  -H "Authorization: Bearer $HEIDI_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "What did we promise Acme in the security review?"}'

{
  "answer": "SSO by Q3, pen-test reports every quarter, and EU
             data residency at renewal. Source: security-review.md,
             call 03-14."
}

# One command. Your coding agent gets your company's memory.
claude mcp add --transport http heidi https://mcp.falkster.ai/mcp \
  --header "Authorization: Bearer $HEIDI_API_TOKEN"

import { Heidi } from "@falkster/heidi";

const heidi = new Heidi({
  baseUrl: "https://mcp.falkster.ai",
  token: process.env.HEIDI_API_TOKEN!,
});

const answer = await heidi.ask(
  "What did we promise Acme in the security review?"
);

const hits = await heidi.memory.search("EU data residency");
// every hit: { text, kind, namespace, source_url }

from falkster_heidi import Heidi

heidi = Heidi(
    base_url="https://mcp.falkster.ai",
    token=os.environ["HEIDI_API_TOKEN"],
)

answer = heidi.ask("Who owns the ingestion pipeline?")

heidi.add_fact("Acme renewal signed 2026-07-30, 2-year term.")
# and AsyncHeidi for your async stack