# fact.guide — Full MCP Server Documentation

> Complete tool and resource reference for the fact.guide MCP server.
> This server provides distributed fact verification with Capacitor synthesis.

## Connection
- Endpoint: https://fact-mcp.forkjoin.ai/mcp
- Protocol: MCP 2024-11-05
- Transport: HTTP POST (JSON-RPC 2.0)
- Authentication: Bearer token or UCAN DID

## Discovery
- Short docs: https://fact-mcp.forkjoin.ai/llms.txt
- Skills: https://fact-mcp.forkjoin.ai/skills/index.json
- Agent manifest: https://fact-mcp.forkjoin.ai/agents.json
- MCP well-known: https://fact-mcp.forkjoin.ai/.well-known/mcp.json
- MCP marketing manifest: https://fact-mcp.forkjoin.ai/.well-known/mcp-manifest.json
- Agent card: https://fact-mcp.forkjoin.ai/.well-known/agent-card.json

## Capacitor Integration
The Capacitor is the intelligent container that wraps facts with:
- DualIndex (Amygdala + Hippocampus) block indexing
- ContentKnapsack layout optimization
- Multi-projection rendering (text, audio, spatial, reading, tiling, hypercube)
- ESI-powered fact enrichment and personalization
- Collaborative editing via CRDT

## REST API Endpoints
- GET /api/synthesis?factId=<id> — Fetch a fact synthesis document
- GET /api/capacitor?factId=<id> — Fetch Capacitor state for a fact
- GET /health — Health check

## Tools (49)

### fact_create
Create a new atomic fact (claim, observation, measurement, inference, or axiom) in the fact graph.
  Parameters:
    - raw_text: The raw text of the fact/claim to create
    - parent_id: Optional parent fact ID for hierarchical nesting
    - fact_type: Type of fact
  Required: raw_text

### fact_update
Update an existing fact node (state, text, type, etc.).
  Parameters:
    - fact_id: ID of the fact to update
    - raw_text: Updated raw text
    - normalized_claim: Normalized claim statement
    - state: New state
    - fact_type: Updated type
  Required: fact_id

### fact_delete
Soft-delete a fact node.
  Parameters:
    - fact_id: ID of the fact to delete
  Required: fact_id

### fact_graph
Get the complete fact graph for the current space, including nodes, links, and verification metrics.
  Parameters:
    - root_id: Optional root node to start from
    - depth: Max depth to traverse (default: 10)
  Required: none

### fact_search
Search for facts by text content or normalized claim.
  Parameters:
    - query: Search query
    - limit: Max results (default: 20)
    - state: Filter by state (raw, pending, verified, refuted, etc.)
  Required: query

### fact_decompose
Use LLM to decompose a compound statement into atomic, independently verifiable claims.
  Parameters:
    - text: Compound statement to decompose
    - fact_id: Or provide an existing fact ID to decompose
  Required: none

### fact_enrich
Use LLM to enrich a fact with normalized claim, suggested sources, and risk assessment.
  Parameters:
    - fact_id: ID of the fact to enrich
  Required: fact_id

### fact_round_create
Start a new commit-reveal verification round for a fact. Validators bond EDGEWORK tokens to participate.
  Parameters:
    - fact_id: ID of the fact to verify
    - required_bond_wei: Required EDGEWORK bond in wei
  Required: fact_id

### fact_round_commit
Submit a blinded commit (hash of vote + salt) to a verification round.
  Parameters:
    - round_id: Verification round ID
    - validator_id: Validator identifier
    - commit_hash: SHA-256 hash of (verdict || salt || roundId)
    - bond_tx_hash: On-chain bond transaction hash
  Required: round_id, validator_id, commit_hash

### fact_round_reveal
Reveal a previously committed vote with the salt and reasoning.
  Parameters:
    - round_id: Verification round ID
    - vote_id: Vote ID from commit phase
    - verdict: The verdict
    - confidence: Confidence 0.0-1.0
    - salt: Random salt used in commit hash
    - reasoning: Reasoning for the verdict
  Required: round_id, vote_id, verdict, salt

### fact_round_settle
Settle a verification round after the reveal phase, computing consensus.
  Parameters:
    - round_id: Verification round ID to settle
  Required: round_id

### fact_round_dispute
File a dispute against a settled verification round within the dispute window.
  Parameters:
    - round_id: Settled round ID to dispute
    - disputer_id: Disputer identifier
    - reason: Reason for dispute
    - appeal_bond_tx_hash: Appeal bond transaction hash
  Required: round_id, disputer_id, reason

### fact_rounds_list
List verification rounds, optionally filtered by phase or fact.
  Parameters:
    - phase: Filter by phase (commit, reveal, settled, disputed)
    - fact_id: Filter by fact ID
  Required: none

### fact_entropy
Get clarity and confidence metrics for the fact space.
  Parameters:
    (no parameters)
  Required: none

### fact_anchor_status
Get Merkle batch anchoring status on Optimism L2.
  Parameters:
    (no parameters)
  Required: none

### fact_anchor_trigger
Manually trigger a Merkle batch anchoring cycle.
  Parameters:
    (no parameters)
  Required: none

### fact_reputation
Get the validator reputation leaderboard.
  Parameters:
    (no parameters)
  Required: none

### fact_sync_pull
Pull new operations since a given clock for CRDT sync.
  Parameters:
    - since: Sync clock timestamp to pull changes after
  Required: none

### fact_presence_heartbeat
Send a presence heartbeat to keep your peer visible in the space.
  Parameters:
    - peer_id: Your peer identifier
    - display_name: Display name
  Required: peer_id

### fact_presence_list
List currently active peers in the fact space.
  Parameters:
    (no parameters)
  Required: none

### fact_synthesize
Generate a full Capacitor synthesis for a fact, including Capacitor blocks, evidence chains, and verification status. Returns a structured synthesis document.
  Parameters:
    - fact_id: ID of the fact to synthesize. If omitted, synthesizes the root of the space.
  Required: none

### fact_synthesize_batch
Batch-synthesize multiple facts into Capacitor documents. Returns an array of synthesis documents.
  Parameters:
    - fact_ids: Array of fact IDs to synthesize.
    - max_depth: Max depth of sub-fact inclusion per synthesis (default: 2).
  Required: fact_ids

### fact_enrich_esi
Trigger ESI-powered LLM enrichment for a fact: normalizes the claim, generates analysis, suggests sources, and assesses risk. Optionally writes the results back to the fact graph.
  Parameters:
    - fact_id: ID of the fact to enrich via ESI.
    - writeback: If true, persist the ESI results back to the fact graph (default: false).
  Required: fact_id

### fact_build_enrichment_prompt
Preview the LLM prompt that would be used for ESI enrichment without executing it. Useful for debugging or customizing the enrichment pipeline.
  Parameters:
    - fact_id: Fact ID to build the prompt for.
  Required: fact_id

### fact_capacitor_state
Get the current Capacitor state for a fact — blocks, projection, layout metadata. This is the data needed to render the fact via the Capacitor container.
  Parameters:
    - fact_id: Fact ID to get Capacitor state for.
  Required: fact_id

### fact_capacitor_project
Switch the active projection surface for a fact — text, audio, spatial, or reading.
  Parameters:
    - fact_id: Fact ID.
    - projection: Target projection surface.
  Required: fact_id, projection

### fact_capacitor_blocks
Get the raw Capacitor blocks for a fact — the building blocks for layout solving.
  Parameters:
    - fact_id: Fact ID.
  Required: fact_id

### fact_build_evidence_chains
Analyze the fact graph to discover implicit evidence chains — links between claims, supporting observations, and conflicting inferences.
  Parameters:
    - root_id: Root fact ID to start from. If omitted, analyzes the entire space.
    - max_chain_length: Maximum chain length to consider (default: 6).
    - apply: When true, creates the discovered links in the graph.
  Required: none

### fact_semantic_search
Search the fact graph using semantic similarity (embedding-based). Returns facts ranked by vector distance rather than keyword match.
  Parameters:
    - query: Natural language search query.
    - limit: Max results (default: 10).
    - threshold: Minimum similarity threshold 0.0-1.0 (default: 0.3).
  Required: query

### fact_flow_decompose
LLM-powered decomposition that breaks a compound claim into atomic sub-facts and creates them as children in the fact graph.
  Parameters:
    - fact_id: Existing fact ID to decompose.
    - text: Or provide raw text to decompose.
    - auto_create: When true, automatically creates sub-facts (default: true).
  Required: none

### fact_collab_start
Start a collaborative editing session for a fact via the Collab Durable Object. Returns a session key and connection details.
  Parameters:
    - fact_id: Fact ID to collaborate on.
    - actor_id: Actor/user identifier.
    - display_name: Display name for the collaborator.
  Required: fact_id, actor_id

### fact_collab_join
Join an existing collaborative session. Returns the current session state and participant list.
  Parameters:
    - session_key: Session key from fact_collab_start.
    - actor_id: Actor/user identifier.
    - display_name: Display name.
  Required: session_key, actor_id

### fact_render_claim_review
Generate a ClaimReview JSON-LD structured data object for a fact. Useful for SEO and fact-checking protocol compliance.
  Parameters:
    - fact_id: Fact ID to generate ClaimReview for.
  Required: fact_id

### fact_render_synthesis
Render a fact synthesis as a Markdown document suitable for display or export.
  Parameters:
    - fact_id: Fact ID to render.
    - format: Output format (default: markdown).
  Required: fact_id

### fact_synthesize_with_esi
Synthesize a fact using the Aeon ESI processor pipeline. Uses the configured EdgeWorkersESIProcessor with KV caching, tier enforcement, and telemetry instead of raw inference calls.
  Parameters:
    - fact_id: ID of the fact to synthesize with ESI.
    - writeback: If true, persist ESI results back to the fact graph (default: false).
  Required: fact_id

### fact_memory_push
Push a fact-linked memory entry into a Fractal Memory stack.
  Parameters:
    - stack_id: string
    - fact_id: string
    - entry_id: string
    - title: string
    - text: string
    - actor_id: string
    - op_id: string
    - hlc: string
    - zk_commitment: string
  Required: stack_id, fact_id

### fact_memory_pop
Pop an entry from a Fractal Memory stack.
  Parameters:
    - stack_id: string
    - target_entry_id: string
    - actor_id: string
    - op_id: string
    - hlc: string
  Required: stack_id

### fact_memory_peek
Peek the top of a Fractal Memory stack.
  Parameters:
    - stack_id: string
    - limit: number
  Required: stack_id

### fact_memory_pin
Pin a Fractal Memory entry.
  Parameters:
    - entry_id: string
    - actor_id: string
    - op_id: string
    - hlc: string
  Required: entry_id

### fact_memory_unpin
Unpin a Fractal Memory entry.
  Parameters:
    - entry_id: string
    - actor_id: string
    - op_id: string
    - hlc: string
  Required: entry_id

### fact_memory_get_pool
Read the Fractal Memory pool.
  Parameters:
    - limit: number
  Required: none

### fact_memory_get_recent
Read recent Fractal Memory entries.
  Parameters:
    - limit: number
  Required: none

### fact_memory_get_pinned
Read pinned Fractal Memory entries.
  Parameters:
    - limit: number
  Required: none

### fact_memory_get_feed
Read combined Fractal Memory feed.
  Parameters:
    - cursor: string
    - include: string
    - limit: number
  Required: none

### fact_memory_link
Create a Fractal Memory relation edge.
  Parameters:
    - from_entry_id: string
    - to_entry_id: string
    - relation_type: string
    - actor_id: string
    - op_id: string
    - hlc: string
  Required: from_entry_id, to_entry_id, relation_type

### fact_memory_get_related
Read Fractal Memory relations.
  Parameters:
    - entry_id: string
    - relation_type: string
    - limit: number
  Required: none

### fact_memory_join_collab_stack
Join a Fractal Memory collaborative stack.
  Parameters:
    - stack_id: string
    - actor_id: string
    - display_name: string
  Required: stack_id, actor_id

### fact_memory_get_collab_stack
Read Fractal Memory collaborative stack snapshot.
  Parameters:
    - stack_id: string
  Required: stack_id

### fact_memory_leave_collab_stack
Leave a Fractal Memory collaborative stack.
  Parameters:
    - stack_id: string
    - actor_id: string
  Required: stack_id, actor_id

## Resources (12)

### Fact Graph
  URI: fact://spaces/default/graph
  The complete fact graph for this space including nodes, links, and metrics.
  Type: application/json

### Clarity Metrics
  URI: fact://spaces/default/entropy
  Clarity and confidence metrics for the fact space.
  Type: application/json

### Verification Rounds
  URI: fact://spaces/default/rounds
  Active and recent verification rounds with commit-reveal status.
  Type: application/json

### Validator Reputation
  URI: fact://spaces/default/reputation
  Leaderboard of validator reputation scores.
  Type: application/json

### Anchoring Status
  URI: fact://spaces/default/anchoring
  Merkle batch anchoring status and recent batches.
  Type: application/json

### Active Peers
  URI: fact://spaces/default/presence
  Currently active peers in this fact space.
  Type: application/json

### Capacitor State
  URI: fact://spaces/default/capacitor
  Current Capacitor container state — blocks, projection, layout metadata.
  Type: application/json

### Available Projections
  URI: fact://spaces/default/projections
  List of available projection surfaces (text, audio, spatial, reading, tiling, hypercube).
  Type: application/json

### Fact Synthesis
  URI: fact://spaces/default/synthesis
  Complete synthesis document for the space root, including Capacitor blocks.
  Type: application/json

### Fact Space Metadata
  URI: fact://meta
  Information about this fact-checking space and verification protocol.
  Type: application/json

### Sign-In & Access
  URI: fact://auth
  How to sign in and access personal fact spaces securely.
  Type: application/json

### Capacitor Reference
  URI: fact://capacitor
  Reference documentation for the Capacitor intelligent container interface.
  Type: application/json