Marlin primary sources — 2026-04-30

Project status in this pass

• No single current canonical Marlin whitepaper or litepaper surfaced.
• The clearest current source of truth is the official site, Oyster docs, and the first-party Oyster monorepo.

Primary-source snapshots

1) Homepage

• URL: https://www.marlin.org/
• Key excerpt:
  • The homepage title says “Scalable coprocessors for decentralized compute.”
  • The meta description says Marlin is a verifiable computing protocol featuring TEE- and ZK-based coprocessors to delegate complex workloads over a decentralized cloud.
  • The visible homepage copy emphasizes fast development cycles, private inputs through confidential computing, cheap verification using ZK proofs or TEE attestations, and composable compute outputs.

2) Homepage FAQ / architecture hints

• URL: https://www.marlin.org/
• Key excerpt:
  • The embedded FAQ says Oyster is a Marlin sub-network specializing in TEE-based coprocessors.
  • It says instances can be rented individually or tasks can be delegated among a pool of nodes in a serverless way.
  • It says Oyster nodes can be used through both smart-contract calls and Web2 APIs, and that auditors routinely query nodes for uptime guarantees.
  • It frames Oyster as permissionless for contributors, builders, and infrastructure providers.

3) Build with Confidential VMs docs

• URL: https://docs.marlin.org/oyster/build-cvm/
• Key excerpt:
  • Says Confidential VMs enable secure and private computation in untrusted environments using hardware-based encryption and secure enclaves.
  • Says Oyster CVM tooling is meant to create secure environments that preserve data privacy while still allowing complex computation.
  • The same page exposes first-class concepts including trusted execution environments, remote attestations, reproducible builds, networking, secure networking, and key-management systems.
  • It also links into tutorials, quickstarts, guides, and examples, reinforcing that this is a real builder platform.

4) Build with Serverless docs

• URL: https://docs.marlin.org/oyster/build-serverless/
• Key excerpt:
  • Says Oyster Serverless lets users execute functions without renting or managing instances directly and pay only for execution time.
  • Says requests can be submitted via smart contracts or Web2 APIs.
  • Says the protocol also supports periodic subscriptions through smart contracts.
  • This is one of the clearest signals that Marlin spans both onchain and offchain request models.

5) Oyster monorepo README

• URL: https://github.com/marlinprotocol/oyster-monorepo/blob/master/README.md
• Key excerpt:
  • Describes the repo as the monorepo for the Oyster platform.
  • Lists a large set of components: attestation servers and verifiers, contracts and KMS-related contracts, indexers, enclave images, initialization utilities, kernels, KMS creator/root/derive components, operator control planes and setup helpers, market indexers, Rust/TypeScript/Go SDKs, Docker-enclave support, serverless executor and gateway components, secret store, and transaction manager.
  • This repo structure makes clear that Marlin is operating a deep infrastructure stack, not only publishing a thin protocol wrapper.

6) Site navigation and official surface map

• URL: https://www.marlin.org/
• Key excerpt:
  • The homepage links out to Oyster, Kalypso, docs, Relay, the Hub/ecosystem, the research forum, and GitHub.
  • That first-party link graph matters because it shows the Marlin brand sits over several adjacent products and protocol surfaces rather than a single isolated network.

Working conclusion

Marlin is best cataloged as verifiable compute and confidential-execution infrastructure with multiple operating layers: confidential VMs, serverless execution, attestation and verification, KMS, operator tooling, contracts, and SDKs. The current homepage is more compute/AI-forward than some older Marlin narratives, but the docs and monorepo make clear the real operating surface is a much broader compute-control-plane stack.