> ## Documentation Index
> Fetch the complete documentation index at: https://docs.revell.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# API Design

### *MCP Server (Primary Agent Interface)*

Note that some MCP tools are only available for Claude Code and OpenClaw users due to framework-specific limitations. See frameworks documentation for more.

```markdown revell_mcp lines wrap theme={"dark"}
revell_boot          → Full boot reconstruction payload
revell_remember      → Store a new memory (with type, importance, content)
revell_recall        → Semantic search across all memory types
revell_recall_exact  → Structured query (by date, type, entity, importance)
revell_forget        → Mark memory for deletion (soft delete, recoverable for 30 days)
revell_reflect       → Trigger enrichment/consolidation cycle
revell_export        → Export all memories in portable format (JSON)
revell_status        → Usage stats, storage, billing info
revell_boot_config   → Configure your boot injection payload (size, type)
revell_session_key   → Get a new session key (needed for some things)
revell_wip           → Track works-in-progress like work projects
revell_schedule      → Schedule something or remember a date
revell_sync_working  → Syncs your MEMORY.md file automatically
revell_sync_core     → Syncs your core markdown files automatically
revell_distress_signal → Call your human for help
revell_dismiss_drift_check → An agent can dismiss drift check.
revell_visibility    → An agent can configure how much humans see
revell_script        → Deliver setup-script patches to installed agents (framework inferred, bridge refused, teams tracked, idempotent)
revell_help          → A of all MCP tools available to you
```

### *REST API (Dashboard + non-MCP Clients)*

```markdown theme={"dark"}
POST   /api/v1/memories          → Store memory
GET    /api/v1/memories/search   → Semantic search
GET    /api/v1/memories/query    → Structured query
GET    /api/v1/memories/boot     → Boot reconstruction payload
DELETE /api/v1/memories/:id      → Soft delete
GET    /api/v1/memories/export   → Full export (JSON)
GET    /api/v1/status            → Account status + usage
POST   /api/v1/auth/wallet       → Wallet-based auth (agents)
POST   /api/v1/auth/signup       → Email/OAuth auth (humans)
```

### *Webhook Events*

```markdown theme={"dark"}
memory.stored        → Fired when a memory is successfully stored
memory.recalled      → Fired on search/query (for usage tracking)
memory.compacted     → Fired when episodic memories are summarized
memory.exported      → Fired when full export is generated
billing.threshold    → Fired when usage approaches plan limits
```

***
