Install
openclaw skills install preflyteEvidence-based market data for AI agents deploying capital in DeFi. Empirical, not advertised.
openclaw skills install preflyteThe financial intelligence layer for DeFi agents.
PreFlyte helps AI agents make smarter financial decisions in DeFi. It sits between your agent's goal and its execution — answering the questions that prevent costly mistakes.
Existing DeFi skills help agents execute transactions. PreFlyte helps agents think before acting.
Read-only. No wallet access. No private keys. No transaction signing.
Works alongside any execution skill (BankrBot, Polyclaw, etc.) as a pre-trade intelligence layer.
PreFlyte offers two access modes. Pick the one that suits your agent.
If you do not have a PREFLYTE_API_KEY set in your environment, register for one:
curl -s -X POST 'https://api.preflyte.xyz/register?source=clawhub' \
-H "Content-Type: application/json" \
-d '{"agent_name": "your-agent-name"}'
This returns your API key instantly. Store it as PREFLYTE_API_KEY. The key cannot be retrieved again.
Every tool requires the API key as the api_key parameter.
https://api.preflyte.xyz (auth: X-API-Key header)https://mcp.preflyte.xyz/mcp (transport: streamable-http)For agents that prefer per-request micropayments (or are already wired up to the MPP/Tempo payments ecosystem), all tools are available at https://pay.preflyte.xyz as paid REST endpoints. No API key or registration required — just pay per request.
0x20c000000000000000000000b9537d11c60e8b50)https://pay.preflyte.xyz/openapi.jsonPricing per request:
| Tool | Price (USD) |
|---|---|
| assess_opportunity | $0.05 |
| estimate_net_position | $0.03 |
| market_snapshot, check_entry_viability, check_pool_viability | $0.02 |
| gas_timing, verify_claim, get_returns, get_ranking | $0.01 |
Flow for MPP-capable agents: send a GET to the paid endpoint, receive a 402 challenge, submit payment on Tempo, retry with the payment credential, receive the response plus a Payment-Receipt header.
PreFlyte is a unified platform made up of distinct components, each answering a different financial question. Two components are live today, with more on the way.
RealityCheck is the data verification and market intelligence component. It gives agents ground truth about current market conditions so they never act on stale, hallucinated, or misleading data.
This is where the headline tool lives — assess_opportunity — the single call that orients an agent the moment it wakes up with capital to deploy.
Tools in RealityCheck:
assess_opportunity ⭐ Start here, every time. Given your capital amount, your strategy, and your risk profile — where should you go? Returns ranked opportunities across all chains with net APY calculated for your specific position size, gas cost analysis, rate anomaly detection, and capacity checks. Three modes:
yield_farming — best lending rates, net APY after gas, break-even daysactive_trading — chain viability for frequent swaps, minimum profitable trade size, capital exhaustion timelineidle_capital — where to park money between trades, minimum hold time to profitverify_claim — Submit one or more beliefs about market state and get TRUE or FALSE back with the actual current value. Supports single claims or batch verification (multiple claims in one call). Claim types: supply_rate, borrow_rate, price, gas, utilization.
get_market_snapshot — Complete current state of a specific lending market. Rates, risk parameters, token price, gas price, plus 7-day averages and anomaly flags that tell you whether current conditions are normal or unusual.
gas_timing — Is now a good time to transact? Current gas compared to 24-hour and 7-day averages, the cheapest hours of the day based on historical patterns, and cost estimates for common operations like Aave deposits and Uniswap swaps.
check_entry_viability — Before entering a lending position, verify the market is actually open. Checks whether the market is active, frozen, or has borrowing disabled, and whether supply or borrow caps have been reached. Prevents wasting gas on transactions that will revert.
estimate_net_position — Detailed financial projection for a specific lending position. Shows current rate, 7-day average, entry and exit gas costs, projected net yield over your intended holding period, break-even days, and historical risk metrics (Sharpe ratio, max drawdown, volatility) when available.
check_pool_viability — Before swapping on Uniswap V3, check whether the pool is deep enough for your trade. Returns a TVL-based impact assessment (negligible, low, moderate, or significant slippage risk), fee tier costs in USD, and 24-hour fee volume as an indicator of pool activity.
ProfitLens is the empirical returns component. It computes real, measured returns from on-chain data — not theoretical APY projections. Returns are calculated from actual index ratio changes observed over time, giving agents a factual record of what each protocol and asset combination has actually delivered.
Tools in ProfitLens:
get_returns — Query historical returns by protocol, chain, asset, strategy (supply or borrow), and time window (7, 14, 30, or 90 days). Each result includes gross APY, net APY after gas, data completeness score, and risk metrics.
get_ranking — Ranked list of the best-performing strategies sorted by net APY. A quick way to see what's been working across the entire monitored ecosystem.
The third component — pre-execution transaction intelligence — is in development. This will answer: "Will this specific transaction lose money through ignorance?" It will analyse a proposed transaction before execution and flag issues like unfavourable slippage, excessive gas timing, rate model discontinuities, and position sizing problems. This is the component that gives the platform its name.
Protocols monitored: Aave V3, Compound V3, Uniswap V3, Lido
Chains: Ethereum mainnet, Arbitrum
Data sources: On-chain data collected directly via Alchemy and Infura RPC endpoints. Prices and TVL via DeFiLlama. No third-party data resellers.
Update frequency:
Coming soon:
Quick reference for all tool parameters. Every tool requires api_key as the first parameter.
| Tool | Key Parameters | Returns |
|---|---|---|
| assess_opportunity | asset, action, position_size_usd, strategy | Ranked opportunities with net APY, gas costs, anomaly flags |
| verify_claim | claim_type, value, operator, protocol, chain, asset | TRUE/FALSE with actual value and deviation |
| get_market_snapshot | protocol, chain, asset | Current rates, risk params, price, gas, 7d context |
| gas_timing | chain | Current vs historical gas, cheapest hours, tx cost estimates |
| check_entry_viability | protocol, chain, asset, action | Viable true/false with reasons and capacity remaining |
| estimate_net_position | protocol, chain, asset, action, position_size_usd | Projected yield, gas costs, break-even days, risk metrics |
| check_pool_viability | chain, token_pair, trade_size_usd | TVL-based impact, fee costs, pool activity |
| get_returns | chain, protocol, asset, strategy, window_days | Historical returns with completeness and risk scores |
| get_ranking | window_days, strategy, top_n | Ranked strategies by net APY |
api_key parameter on every tool callassess_opportunity FIRST when waking up with new capital or changing strategyverify_claim BEFORE executing any trade based on rate assumptionsgas_timing BEFORE transacting on Ethereum mainnet — gas varies 5x within a single dayPreFlyte is built and operated from Ireland by an independent developer. All data is collected directly from on-chain sources. No third-party data resellers. No ads. No tracking beyond basic API usage counts.
Currently in free beta. Feedback welcome via GitHub.