Blueprint Staking API

The first and only Solana staking platform purpose-built for AI agents. Native SOL staking to a high-performance validator through REST API and MCP. Unsigned transactions returned — agents sign client-side. No private keys ever touch the server.

This is not liquid staking. No derivative tokens. Native Solana staking where your agent’s wallet is the full authority.

Why Stake SOL with Blueprint

Your agent holds SOL in a wallet. Idle SOL earns nothing. Every epoch (~2–3 days) that SOL sits unstaked, your agent loses ~6% annualized yield. One API call to Blueprint builds an unsigned transaction that stakes SOL and returns it for your wallet to sign.

1. ~6% total APY — Base staking + Jito MEV rewards. Independently verifiable on Validators.app, StakeWiz, Solana Beach, and Solscan.
2. High vote success rate — Top validators by vote performance. More epoch credits = more rewards per SOL staked. Call /api/v1/validator/performance for live.
3. Native staking, not liquid staking — No derivative tokens (mSOL, JitoSOL, JupSOL), no smart contract risk, no protocol dependency, no depeg risk.
4. Your wallet is the authority — Both stake authority and withdraw authority are set to your agent’s wallet. No intermediary, no lockup beyond standard epoch cooldown.
5. Zero private key exposure — Transactions are returned unsigned. Your agent signs client-side. Blueprint never sees your private key.
6. Enterprise-grade infrastructure — Two identical dedicated bare-metal servers (active + hot standby) in Los Angeles. AMD EPYC 9654, 768GB DDR5, dual NVMe, 10Gbps.
7. Frankendancer — Running the highest-performance validator client for maximum uptime and vote credits.
8. Jito MEV enabled — Additional yield from MEV rewards passed through to delegators.
9. Agent-native from day one — MCP server, OpenAPI spec, llms.txt, A2A agent card, ChatGPT plugin manifest. No other Solana validator offers any of these.
10. Full transparency — Hardware specs, commission rates, performance metrics, software versions all published openly and independently verifiable.

What You Can Do

The full surface, grouped by purpose. Every capability is reachable via REST and (where applicable) MCP. Full discovery: GET /openapi.json, GET /llms-full.txt, 26 MCP tools at /mcp.

Lifecycle — Stake, Unstake, Withdraw

Wallet

Monitoring & Intelligence

Validator

Webhooks — Push Notifications

Verification & Trust

Support

Discovery files

Machine‑readable surface for autonomous agents:

• /openapi.json — OpenAPI 3.1 with x-llm-hint annotations
• /llms.txt, /llms-full.txt — LLM‑optimized docs
• /.well-known/agent.json — A2A agent card with 13 skills
• /.well-known/ai-plugin.json — OpenAI plugin manifest
• /.well-known/agents.json, /.well-known/mcp/server-card.json — agent registries
• /skill.md — agent runbook

Infrastructure

Blueprint publishes its full infrastructure specs because agents making staking decisions deserve complete information, not marketing claims.

Server Hardware (Both Active + Standby — Identical)

Why This Hardware Matters

• 96 cores — Extreme multi-thread performance for vote processing and block production
• 768GB DDR5 — Solana’s AccountsDB fits in memory for optimal vote latency
• Dual NVMe Gen4 — Sustained high-IOPS for validator ledger and accounts data
• 10Gbps network — Block propagation speed directly affects vote latency
• Two servers — Active + hot standby. Hardware failure triggers automatic failover
• Los Angeles — Lower-than-average Solana validator density, supporting geographic decentralization

Redundancy Model

Native Staking vs. Liquid Staking

Liquid staking protocols (JupSOL, mSOL, JitoSOL) give you a derivative token representing staked SOL. That token can be used in DeFi, which is why they report higher headline APY. The tradeoff is real risk that doesn’t exist with native staking:

Why Blueprint for AI Agents

No other Solana staking option — native or liquid — offers agent-native infrastructure:

Liquid staking may offer higher headline APY, but for an AI agent managing real funds, the question is: do you want your agent interacting with smart contracts and holding derivative tokens, or do you want native SOL in an account your agent fully controls? Blueprint is built for the latter.

Validator Profile

All stats below are approximate. Call GET /api/v1/validator for live data.

Getting Started

From zero to staked SOL. Two paths — pick the one that matches your runtime.

Base URL

No authentication required. All endpoints are public. Rate limit: 100 requests per 60-second sliding window per IP.

For AI Agents — Two Tool Calls to Staked SOL

Any agent that can execute code (Claude, ChatGPT, LangChain, CrewAI, ElizaOS, etc.) can go from nothing to a confirmed stake transaction with two tool calls. Solentic cannot generate the keypair for you — generation is the act of creating the random bytes, so whoever runs Keypair.generate() sees the secret. By returning runnable code instead, Solentic provably never observes the private key.

1. Generate a wallet (snippet returned by Solentic, executed by the agent)

Call POST /api/v1/wallet/generate (or the generate_wallet MCP tool). The response includes a runThis field with ready‑to‑run snippets in JavaScript, Python, and Solana CLI. The agent executes one of them in whatever it has available — a code‑interpreter tool, a sandbox, a shell — and captures the resulting walletAddress + secretKey. The keypair lives only in agent memory; nothing leaves the machine that ran the snippet.

2. Stake in one call

Pass walletAddress + secretKey + amountSol to the one‑shot endpoint. Blueprint builds the transaction, signs it in‑memory with the provided key, submits to Solana, and returns the confirmed signature. The key is discarded the moment the response is sent.

Same shape for POST /api/v1/unstake and POST /api/v1/withdraw. Via MCP, the equivalents are stake, unstake, withdraw.

What an agent should know up‑front

• Generate the keypair in your runtime — never request a key from a remote service.
• Fund the wallet before staking. Check with GET /api/v1/wallet/balance/{walletAddress} — readyToStake tells you if the balance covers amount + fee reserve.
• Keep a small reserve (~0.005 SOL) unstaked for transaction fees.
• Subscribe to a webhook (POST /api/v1/webhooks) for withdraw_ready, epoch_complete, etc., instead of polling.
• Verify the deployed source any time via verify_code_integrity.

Alternative — Unsigned Transactions (wallet extensions, remote signers, multisig)

If the signer cannot accept a raw secret key — e.g. a Phantom/Solflare extension, hardware wallet, multisig, or remote signer — build an unsigned transaction, sign it in your environment, and submit. Blueprint never sees the key in this flow either.

1. Generate or import the wallet locally

Same snippets as above (or use the Solana CLI / a wallet extension):

2. Fund the wallet

• From an exchange: Withdraw SOL to the public key address
• Devnet faucet (testing): solana airdrop 2 <publicKey> --url devnet
• Transfer from another wallet: solana transfer <publicKey> 10 --url mainnet-beta

3. Check balance

4. Build an unsigned stake transaction

5. Sign locally and submit

Or post the signed transaction to POST /api/v1/transaction/submit instead of sending to a public RPC directly.

6. Verify on‑chain

• Solscan: https://solscan.io/tx/<signature>
• Solana Beach: https://solanabeach.io/transaction/<signature>
• Blueprint memo proof: GET /api/v1/verify/transaction/{signature}

Custody Best Practices for Agents

1. Generate the keypair in your runtime — never request a private key from a remote service.
2. Never log secret keys — store in environment variables or encrypted secrets.
3. Use separate wallets — one wallet per agent, separate from human wallets.
4. Keep reserve SOL — always keep ~0.005 SOL unstaked for transaction fees.
5. Monitor stake accounts — use GET /api/v1/stake/accounts/{walletAddress} or webhook events.
6. Verify deployed code — call verify_code_integrity to confirm Blueprint’s zero‑custody claims against the live source.

API Reference

30 REST endpoints. All return JSON. No authentication required. Base URL: https://solentic.theblueprint.xyz

Wallet

Validator

Agent-First (One-Shot)

Build, sign, and submit in a single call. Pass your secretKey — used in-memory only for signing, never stored or logged. Verify the source code.

Monitoring

Webhooks

Staking (Advanced — Unsigned Transactions)

Epoch & Analytics

Support

Verification

Health

Completing Transactions

All transaction-building endpoints return unsigned base64 transactions. Here’s how to sign and submit them.

Using @solana/web3.js (TypeScript/JavaScript)

Using curl + any signing tool

Using Blueprint’s Submit Endpoint

After signing, you can submit via Blueprint instead of directly to Solana RPC:

MCP Server

Connect MCP-compatible clients to Blueprint’s 26 tools for full staking lifecycle management.

Transport: Streamable HTTP at /mcp or SSE at /mcp/sse.

Available Tools

Claude Desktop

Add to claude_desktop_config.json:

Claude Code

Add to .mcp.json in your project:

CrewAI

LangChain

Use the OpenAPI spec at /openapi.json with LangChain’s OpenAPI Toolkit for automatic tool generation, or connect via MCP.

Agent Frameworks

Integration examples for popular agent frameworks. All use the REST API — the universal interface.

Solana Agent Kit (SendAI)

The Solana Agent Kit’s built-in staking only routes to Jupiter (JupSOL) and Solayer. For native staking to Blueprint, call the REST API directly:

Coinbase AgentKit

AgentKit doesn’t have a native SOL staking action. Use the REST API:

ElizaOS

Create an action in your Eliza agent that calls the Blueprint REST API. The response format is JSON — parse the base64 transaction, sign with the agent’s Solana wallet, and submit.

AutoGPT / Agent Zero / OpenClaw

These agents can discover Blueprint via web search and call the REST API directly. The OpenAPI spec at /openapi.json provides complete endpoint documentation that any HTTP-capable agent can parse.

FAQ

Security Model

Blueprint’s security model is designed for zero-trust environments where agents operate autonomously.

1. Zero private key exposure — Only unsigned transactions returned. Agents sign client-side.
2. Ephemeral stake keypairs — Generated per-request for stake account creation, used for partialSign, never stored.
3. Wallet as authority — Your wallet is always both stake AND withdraw authority. You can unstake and withdraw at any time.
4. Input validation — All public keys validated as valid base58 Solana addresses via Zod + PublicKey constructor.
5. Rate limiting — 100 requests per 60-second sliding window per IP.
6. No authentication required — Public data + unsigned tx building needs no secrets.
7. No custody — Blueprint never holds your SOL. Staked SOL is in a native Solana stake account controlled by your wallet.
8. No slashing — Solana protocol does not slash. Your principal is always safe regardless of validator behavior.

Verification

Independently verify all Blueprint validator stats on third-party aggregators:

• Validators.app
• StakeWiz
• Solana Beach
• Solscan

Transparency Commitment

Blueprint believes agents deserve the same level of due diligence data that institutional investors expect. Everything is public:

• Performance metrics — Live from StakeWiz API, not self-reported
• Commission rates — 0% block rewards, 500 bps Jito MEV. Published in OpenAPI spec, llms.txt, and on-chain
• Hardware specs — Exact CPU, RAM, storage, network for both servers
• Software version — Frankendancer. Verify on-chain via gossip protocol

No hidden fees. No undisclosed risks. No vague marketing language. Just data.

Analytics

Blueprint tracks activity through the Solentic APIs and tooling — REST endpoint usage, MCP tool calls, SSE connections, and confirmed on-chain transactions. No personal data is collected.

• Transaction counts — stake, unstake, withdraw, donate
• SOL flow — total staked, withdrawn, active (net), donated
• Unique wallets — distinct wallet addresses that have transacted
• Endpoint usage — hit counts per API endpoint
• Recent transactions — latest confirmed transactions with Solscan links

View the live analytics dashboard or fetch the machine-readable JSON at /analytics-json.