← ObservatoryGuidesMCP Access

Guide — Machine Interface

Using the Observatory with AI Assistants

Ask questions across the entire Observatory instead of reading records one at a time. Connect an AI assistant and query claim status, lifecycle history, and cross-record patterns in plain language.

Endpoint
mcp.faultlinewatch.com/mcp
Status
Live
Access
Read-only
Compatible clients
Codex, Claude Desktop, MCP Inspector, and other MCP-compatible clients
Current capabilities
Claim status lookup · FCIF lifecycle · Case listing · Observatory metadata

Connecting your client

Add the Observatory endpoint to your MCP-compatible client. Configuration format varies by client — examples for common clients are below.

config.toml
[mcp_servers.faultline]
url = "https://mcp.faultlinewatch.com/mcp"

Access is public and read-only. No API key, account, or authentication step is currently required. Most users can connect directly using the endpoint shown above. Reasonable rate limits may apply to protect service availability, particularly for large cross-record queries or automated loops.

Once connected, your client will discover the Observatory tools automatically. Current tools expose capabilities such as record listing, claim lookup, lifecycle inspection, and Observatory metadata retrieval. The exact tool names displayed will depend on the MCP client you are using.

After adding the endpoint, restart your client, confirm that the Faultline Observatory tools appear in the available tool list, and begin a new chat session. Start with a simple query to verify the connection before moving to broader corpus-level analysis.

Try

  • What is the current FCIF status of LK-99?
  • Which claims are currently invalidated?
  • Explain the FCIF lifecycle stages.
  • List all tracked cases in Quantum Computing.

Cross-record analysis

Most visitors use the website to read individual Frontier Records. The MCP connection enables a different mode of interaction. Instead of opening one record at a time, an AI assistant can query Observatory tools directly and analyse information across the corpus.

For example:

  • Which claims have accumulated the strongest contradictory evidence?
  • Which records have remained stable for long periods without reassessment?
  • Which programmes contain the largest number of unresolved open questions?
  • How does evidentiary behaviour differ between programmes?

Questions like these require looking across many records simultaneously. The MCP connection makes that possible.

What the records do not do

not predictiveThe Observatory records evidence and assessment history. It does not predict outcomes. A claim's current state describes its present evidentiary position, not its future.
not exhaustiveThe Observatory is selective rather than exhaustive. Absence from the corpus should not be interpreted as evidence for or against a claim.
read-onlyThe public machine interface is read-only. Neither AI assistants nor connected clients can modify Observatory records through the MCP connection.