Fastest fetch in DeFi
No RPC dependency. Solve 100% locally. Rust-native rewrites of popular DEXs and in-memory revm simulations for the rest. Scale from 1 to 100 parallel workers.
Fynd
10-50 ms
Aggregator API
250-750 ms
Intent Auction
1-50 s
Nodeless
All DEX state streams to you in real-time from Tycho websockets, 10x faster than querying a node.
Rust native
Popular DEXs reimplemented in Rust: 900 nanoseconds per pool simulation. No RPC overhead, just raw performance.
Unlimited scale
Auto-scale from 1 to 100 router workers running in parallel, async from the HTTP runtime. Up to 500 RPS.
Zero external calls
No third-party APIs or RPCs at runtime. Everything resolves locally. New pools and tokens appear seconds after deployment.
Nothing gets past Fynd
More than half of Ethereum solvers rely on Tycho, the indexer behind Fynd. It covers 40+ protocols, 180,000+ pools, and 95,000+ tokens across Ethereum, Base, and Unichain.
Chains
Teach Fynd new tricks
Modify any component: algorithms, pool selection, solving strategy. Make Fynd yours.
Open source, with support. Build whatever you want with Fynd; it's MIT licensed. Count on our team for guidance.
Join us on Telegram →
Settle more reliably
Restrict routing to safe tokens, deep pools, specific DEXs, or exclusively RFQ quotes. Optimize for fills over price to reduce failed transactions.
Find better routes
Three solving modes: fast quote (<15ms), safe route (<50ms), and exhaustive optimal search (<5s). Run them in parallel and auto-select the best result.
Trust your quotes
No approximations. Exact onchain state with verified DEX logic. Every quote is precise, verifiable, and ready to execute onchain
Focus on your pools
Build a custom router for your DEX, your tokens, or your hook. Include your own private liquidity in trade paths. Route long-tail tokens seconds after they go live.
Best in show
See the difference when you own your routing stack
Fynd
DEX Aggregator
Intent Auction
Response time
10-50ms
250–750ms
1–50 seconds
Rate limits
Unlimited (self-hosted)
Strict limits
-
Trust model
Trustless (verify locally)
Fully trusted
Fully trusted
Token coverage
Every token & pool
Established tokens
Major tokens only
Customization
Fully open source
Limited
Not customizable
API cost
None (self-hosted)
Subscription + usage fees
PI + frontend fees
Execution reliability
High (configurable)
Limited
Limited
Routing quality
Good (improving)
Very good
Very good