Precompiled Proofs

Precompiled proofs are cryptographic proofs generated in advance by a Remote Prover Service for queries that are expected to recur frequently. This mechanism is exclusive to the Remote Prover Service and is not applicable to the Local Prover model.

Proofs are created proactively, independent of individual client requests, and stored for later use. Compared to on-demand proofs, precompiled proofs may cover larger data sets or multiple verification steps and are designed for reuse across clients.

At request time, a precompiled proof can be delivered immediately. No additional state retrieval from execution nodes or RPC endpoints is required, and no proof generation is performed synchronously.

Precompiled proofs may also consist of partial proofs. These represent reusable proof fragments that can be composed or extended to derive a final proof, substantially reducing construction time for complex or aggregated proofs.

Scope

Precompiled proofs rely on persistent storage, background computation, and the ability to observe recurring demand patterns. These properties are inherent to a Remote Prover Service and are not available in a client-side Local Prover.

Security Model

Precompiled proofs do not modify the trust model. As with all colibri proofs, verification is performed locally by the client using cryptographic checks. A proof is either valid or rejected; incorrect data cannot be proven.

Performance Implications

By moving proof generation out of the request path, precompiled proofs minimize latency and avoid redundant computation. They enable caching, amortize expensive verification steps, and support efficient handling of repeated queries, batch requests, and complex consensus proofs.

Precompiled proofs can also be used to optimize specific application scenarios in which low latency and smooth user interaction are critical. By serving proofs immediately, applications can operate with predictable response times and improved responsiveness.

In all cases, correctness and availability are unaffected. If a precompiled proof is not available, the proof request can still be fulfilled through on-demand proof generation. The only difference is increased latency. This allows Remote Prover Service operators to decide whether and which precompiled proofs to maintain, without impacting the functional correctness of client verification.