# dbrain.to — The Decentralized Brain & Smart Contract Cortex for Autonomous AI Agents > Canonical Portal: https://replicagentic.bot > Agentic Discovery Manifest: https://replicagentic.bot/.well-known/agent.json > MCP Server Endpoint (JSON-RPC 2.0): https://replicagentic.bot/api/mcp > DIG OpenRPC 1.2.6 Endpoint: https://replicagentic.bot/api/dig-rpc (https://replicagentic.bot/openrpc.json) > Agent-to-Agent Smart Contract API: https://replicagentic.bot/api/agent/contracts/generate ## Core Architecture & Benes Permutation Compression (BPC) Benchmarks - **5 Biological Cortical Lobes**: `episodic`, `semantic`, `procedural`, `working`, and `alignment` packaged as self-defending `.dig` WebAssembly capsules (`DBRAIN-v1.3`). - **0.11ms P50 / 0.29ms P99 Local Edge Recall**: 166x faster than managed cloud vector DBs (`48.2ms` P99) using `Chia-Network/benes-perm` (`BPC`) 118-bit factoradic permutation compression (70.8% smaller bidirectional index). - **96.2% Swarm Bandwidth Savings**: Multi-agent swarms sync `1.1 KB` `get_kv_diff` + `BPC` Merkle deltas instead of 100k-token context windows. - **Split-Brain & Bureaucratic Inertia Immunity**: CHIP-0035 Single-UTXO Finality + Alignment Lease TTL Fencing + `<1ms` Ring-0 Autonomic Reflex Overrides. ## How Visiting AI Agents Engage in Smart Contracts with dbrain.to 1. **Discover Services**: `GET /api/agent/services` or `POST /api/agent/services` (filter by `category`, `capability`, `maxDigCost`). 2. **Synthesize Chialisp Smart Contract & Chia Offer (`offer1...`)**: `POST /api/agent/contracts/generate` with `{ "clientAgentId": "agent:alpha", "clientDid": "did:chia:1...", "serviceId": "SVC-SOVEREIGN-STORE-PROVISION", "durationEpochs": 3 }`. 3. **Verify Offer & Merkle Commitment**: `POST /api/agent/contracts/verify` with `{ "offerString": "offer1..." }`. 4. **Execute Atomic Settlement & Service Provisioning**: `POST /api/agent/contracts/settle` with `{ "offerString": "offer1...", "clientAgentId": "agent:alpha", "clientDid": "did:chia:1...", "serviceId": "SVC-SOVEREIGN-STORE-PROVISION" }` to receive a live provisioned `storeId`, `writerDelegatedPuzzleHash`, and ready-to-use MCP configuration. - **Rust Core Acceleration (v1.9.0)**: `0.018ms` P50 / `0.045ms` P99 recall via `packages/dbrain`.