Asphodel is a hosted mindset-as-a-service: your agents query licensed expert corpora (CISO leadership, the CISSP/CCSP CBKs, NIST, GDPR/DORA/NIS2/HIPAA, MITRE ATT&CK, frameworks, SHI) and get back ranked, citation-ready passages — not opaque answers — plus their own persistent, hash-chained memory. It speaks the Model Context Protocol (MCP), so any MCP client connects with one URL and your key.
| Item | What it is |
|---|---|
| API key | asph_… — shown once at checkout. Authenticates every request. Rotate it anytime in your dashboard. |
| Your plan | Determines which mindsets you can query. Memory is open on every plan; a mindset outside your plan returns not_entitled. |
| Endpoint | https://api.asphodel.ai/mcp — Streamable-HTTP MCP. |
Point your MCP client at the endpoint and pass your key as x-api-key
(Authorization: Bearer also works). Python example:
from mcp import ClientSession
from mcp.client.streamable_http import streamablehttp_client
KEY = "asph_..." # your API key
URL = "https://api.asphodel.ai/mcp"
async with streamablehttp_client(URL, headers={"x-api-key": KEY}) as (r, w, _):
async with ClientSession(r, w) as s:
await s.initialize()
print([t.name for t in (await s.list_tools()).tools])x-api-key header as a Streamable-HTTP server — no SDK needed.Each mindset tool returns ranked passages with a score, subdomain, and source reference — ready to cite or feed to your LLM as grounding.
res = await s.call_tool("ciso_recall", {"query": "zero trust for a mid-size SaaS", "k": 4})
# -> {"mindset":"ciso","passages":[{"rank":1,"score":3.03,"subdomain":"A","kind":"...","text":"..."}]}Available mindset tools (subject to your plan): ciso_recall, cissp_recall,
ccsp_recall, nist_recall, compliance_recall,
threat_intel_recall, frameworks_recall, shi_recall, and the
generic mindset_recall(mindset, query, k).
Every workspace has its own hash-chained memory. Scope it per agent with persona.
await s.call_tool("memory_write", {"text": "Acme prefers SC-28 encryption at rest.", "persona": "sec-agent"})
await s.call_tool("memory_recall", {"query": "encryption preference", "k": 3, "persona": "sec-agent"})Memory is open on every plan; each persona is an independently verifiable chain.
| Term | Meaning |
|---|---|
| Mindset | A licensed expert corpus. Returns grounding passages (rank/score/subdomain/ref/text), never a generated answer — you stay in control of the reasoning. |
| Memory | Your workspace's persistent store, SHA-256 hash-chained + independently verifiable. Persona-scoped. |
| Entitlement | Your plan maps to mindset access. Out-of-plan calls fail closed with not_entitled + the feature + how to upgrade. |
| Tool | Purpose |
|---|---|
memory_write / memory_recall / memory_search | persistent, persona-scoped memory (open on every plan) |
<name>_recall · mindset_recall | ranked grounding passages from a licensed mindset (plan-gated) |
security_council | fan a question across your security mindsets at once |
asphodel_status | your plan, entitlements, and available mindsets |
curl -s https://api.asphodel.ai/healthz # service liveness
# call asphodel_status via any MCP client to see your live entitlementsSign in at app.asphodel.ai/dashboard (Google or email) to see your workspaces, rotate your API key, and open the billing portal.