Rollkit primary sources
Reviewed: 2026-05-15 UTC
Original launch thesis / modular-rollup framing
- https://blog.celestia.org/introducing-rollkit-a-modular-rollup-framework/
- Best source for Rollkit’s original self-description.
- Used to confirm that Rollkit was introduced as a modular rollup framework, formerly Rollmint, with an ABCI-compatible client interface and an explicit goal of letting developers deploy rollups across a modular stack.
- Also used to confirm the project’s original validator-elision thesis: developers could rely on an underlying DA layer rather than gathering and compensating a standalone validator set.
- Especially important because it names the major intended control surfaces directly: pluggable execution environments, multiple sequencing modes, different proof schemes, and multiple rollup types.
Current historical status / rebrand continuity
- https://raw.githubusercontent.com/rollkit/.github/main/profile/README.md
- Best concise source for the project’s current status.
- Used to confirm that Rollkit is now Evolve, that the codebase moved under
evstack, and that future development is expected under the newer brand. - Important because it prevents treating Rollkit as a dead-end artifact; it is better understood as the historical predecessor and earlier framing of the current Evolve stack.
Data-availability interface
- https://raw.githubusercontent.com/rollkit/go-da/main/README.md
- Best source for the DA abstraction layer.
- Used to confirm that Rollkit externalized DA through a generic interface with methods for
Get,GetIDs,GetProofs,Commit,Validate, andSubmit, with optional namespace support depending on the backend. - Also useful because it lists multiple DA implementations, making backend substitution a first-class part of the architecture rather than an afterthought.
Sequencing interface
- https://raw.githubusercontent.com/rollkit/go-sequencing/main/README.md
- Best source for the sequencing abstraction layer.
- Used to confirm that Rollkit externalized ordering through a separate interface with methods for transaction submission, next-batch retrieval, and batch verification.
- Also useful because it shows sequencing was intended to be implementation-pluggable, with documented centralized, Astria-connected, proxy, and local/mock paths.
Concrete sequencer implementation
- https://raw.githubusercontent.com/rollkit/centralized-sequencer/main/README.md
- Best source for one concrete sequencing deployment pattern.
- Used to confirm that the reference centralized sequencer runs as a gRPC service and posts ordered rollup transactions to a DA layer.
- Important because it makes sequencer operation into a visible middleware role with its own host, port, namespace, batch-time, and metrics controls.
Execution attachment evidence
- https://raw.githubusercontent.com/rollkit/go-execution-evm/main/README.md
- Best source in this pass for the execution-attachment side of the stack.
- Used to confirm that the Rollkit org later exposed a dedicated EVM execution client implementing the Rollkit execution interface via Ethereum Engine API and standard JSON-RPC.
- Useful mainly because it shows the execution environment was not just a marketing promise in the original launch post; it became an explicit integration boundary.
Main takeaways from this pass
- Rollkit clears the bar because it isolates the older interface-first architecture underneath the current Evolve branding.
- The most reusable mechanism is not
modular rollupin the abstract, but the split between application logic, sequencing middleware, DA adapters, and attached execution environments. - Rollkit is especially useful as a comparison point for Evolve, Sovereign SDK, Commonware, and other chain-launch stacks because it makes validator-elision and interface substitution explicit.
- The rebrand matters analytically: Evolve inherits much of the same architecture, but cataloging Rollkit separately preserves the original public-good and neutrality framing that can otherwise disappear in productized summaries.