Proposal: DamlHat — Unified Canton Developer Toolchain & Local Devnet

DamlHat — Unified Canton Developer Toolchain & Local Devnet

Author: Riccardo Ravaro & Adrian Vide Paris Status: Draft Created: 2026-04-06

Contact: r***@datatamer.ai

Summary

DamlHat is a Hardhat/Foundry-inspired developer toolchain for the Canton Network. A single CLI scaffolds Daml projects, spins up a full multi-node local devnet in Docker, generates typed client SDKs, simulates transactions, and deploys to testnet — with an integrated transaction trace viewer for debugging. The goal is to collapse the current multi-day "infrastructure engineer" onboarding into a single damlhat init && damlhat node up command.

Objective and Scope

The 2026 Canton Developer Experience Survey identified infrastructure complexity as the #1 pain point (41% of respondents) and a unified CLI framework as the most requested tool. DamlHat directly addresses both. Scope:

• A unified CLI (damlhat) covering scaffold, build, test, deploy, and node management.
• A one-command local devnet bundling participant nodes, sequencer, mediator, and a test wallet — reproducible across macOS/Linux/Windows via Docker.
• Typed client SDK codegen for TypeScript and Python from compiled DAR files, removing manual hash/package-ID extraction.
• A transaction trace viewer (CLI + minimal web UI) that decodes Daml transaction trees, exercised choices, and failures — inspired by Tenderly.
• Reference project templates: a token, a simple DEX, and a multi-party workflow.

Out of scope: mainnet validator operation, custom consensus work, non-Daml language frontends.

Technical Approach

• CLI core: TypeScript, published to npm as @damlhat/cli. Plugin architecture so the community can extend commands.
• Local devnet: Docker Compose topology pinned to current Canton releases, with health checks, deterministic party allocation, and a pre-funded faucet. Optional Kind/k3d mode for Kubernetes parity.
• Codegen: Wraps daml codegen and emits ergonomic typed clients with full JSDoc/typing, plus Python pydantic models. Handles Package ID discovery automatically against the JSON API.
• Trace viewer: Parses ledger API transaction trees into a timeline view (exercised choices, created/archived contracts, observers). Ships as damlhat trace <tx-id> plus an optional local web UI served by the CLI.
• Testing: Jest-style test runner that boots ephemeral ledgers per test file, with snapshot-based assertions on contract state.

Architectural Alignment

DamlHat is additive and does not modify the Canton protocol. It consumes the public Ledger API, JSON API, and Canton admin APIs. It complements ongoing work on cantonctl by focusing squarely on the application developer loop (local devnet + typed SDKs + trace debugging) rather than operator tooling. All output artifacts remain compatible with stock Canton tooling — developers can drop DamlHat at any time without lock-in.

Milestones and Deliverables

M1 — CLI core & local devnet (6 weeks)

• damlhat init, damlhat build, damlhat node {up,down,reset}
• Dockerized multi-node devnet with faucet and pre-allocated parties
• Public alpha release on npm, CI for macOS/Linux

M2 — Typed SDK codegen (5 weeks)

• TypeScript + Python codegen from DAR files
• Automatic Package ID resolution helper
• Two worked examples (token, escrow) with end-to-end tests

M3 — Transaction trace viewer (6 weeks)

• damlhat trace CLI command
• Local web UI with timeline visualization, choice decoding, error explanations
• Integration with the test runner for failed-assertion traces

M4 — Templates, docs, community release (5 weeks)

• Three reference templates (token, DEX, multi-party workflow)
• Full documentation site, quickstart video, migration guide from bare Daml SDK
• v1.0.0 tagged release, Apache 2.0 license

Acceptance Criteria

• A new developer with no prior Canton experience can reach a deployed testnet contract in under 30 minutes following the quickstart.
• damlhat node up produces a working multi-node devnet in under 2 minutes on a standard laptop.
• Typed SDK codegen produces zero-warning TypeScript and passes mypy --strict for Python.
• Trace viewer correctly decodes at least 95% of transaction types on a regression corpus of 200+ recorded transactions.
• All deliverables published as open source (Apache 2.0) with CI, test coverage ≥80%, and public documentation.

Funding Request and Milestone Breakdown

Total requested: 150,000 USD equivalent in Canton Coin, milestone-based.

Milestone	Deliverable	Payment
M1	CLI core & local devnet alpha	40,000
M2	Typed SDK codegen	35,000
M3	Transaction trace viewer	40,000
M4	Templates, docs, v1.0 release	35,000

Duration: ~22 weeks from grant approval.

Strategic Alignment

DamlHat is a common-good piece of infrastructure that lowers the barrier to entry for every new Canton builder — directly supporting the Foundation's priority areas of developer tools and reference implementations. It compounds the value of Zenith's EVM onramp by giving native Daml developers the same quality of loop that Hardhat/Foundry users expect, and it turns the most-cited survey pain point into a solved problem.

Team & Background

The project is delivered by a two-person team.

Riccardo Ravaro (lead) — Senior Full Stack Engineer, 10+ years, based in Milan, Italy.

Directly relevant experience:

• Rabobank (2022–2023) — built a trading platform on Corda blockchain (the R3 lineage Canton shares), full front-end and back-end, React/Next.js dashboards, microservices on GCP.
• Main Incubator / Commerzbank — Marco Polo Network (2018–2019) — Corda-based trade finance application used by major banks; Node.js and Python microservices, Kubernetes, CI/CD.
• Pillar, Attrace, Forge, AffiliateCoin (2018–2022) — multiple Ethereum/Solidity smart-contract projects, Web3.js tooling, staking contracts, SDK maintenance.
• DataTamer.ai (2025–present) — senior full-stack, LLM/RAG data architecture, agentic workflows.
• Consistent track record shipping production systems with Kubernetes, CI/CD, PostgreSQL, Docker, and TypeScript — the exact stack DamlHat is built on.

Corda and Canton both originate from the R3 / Daml lineage, so the lead applicant brings direct operational experience with the transaction model, party semantics, and node topology that Canton inherits. This substantially de-risks delivery of the local devnet, trace viewer, and typed SDK work.

Adrian Vide Paris (co-applicant) — Senior Full Stack & DevOps Engineer. Lead full-stack experience across projects in the UK and Spain. Responsible on this project for Kubernetes/Helm/Terraform infrastructure, the Docker-based local devnet, and CI/CD pipelines.

• City: Barcelona, Spain
• Contact: a***@gmail.com
• GitHub: https://github.com/adrianVide
• Current role: Lead Contributor @ DataTamer.ai (same team as the lead applicant)

Work split:

• Riccardo — specification, architecture, Canton/Corda ledger integration, TypeScript libraries and SDK codegen, trace viewer core.
• Adrian — local devnet (Docker/Kubernetes), Helm/Terraform deployment, CI/CD, release engineering, cross-platform packaging.

Contact (lead): r***@datatamer.ai

Co-Marketing

The team will collaborate with the Canton Foundation on visibility and ecosystem promotion, including:

• Launch announcement coordination at v1.0 release
• Technical blog posts on the DataTamer.ai engineering blog and the Canton forum at each milestone (M1–M4)
• X Spaces / AMAs with the Canton Network team to walk developers through the toolchain
• Conference / meetup talks where relevant (e.g. a "Hardhat for Canton" lightning talk)
• Tutorial content — written quickstarts and at least one screencast per milestone, contributed under the Canton documentation umbrella
• Open development — weekly progress posts on the Canton forum during the grant period, so the community can follow along in real time

Notes for Reviewers

• This proposal is intentionally scoped around the application developer inner loop. It does not overlap with operator/validator tooling covered by other proposals such as cantonctl.
• All code will be developed in the open from day one, with weekly progress updates posted to the forum.
• Maintenance commitment: the team will provide bug fixes and Canton version compatibility updates for 12 months post-v1.0 at no additional cost.