dev.fast
/dev/fast builds developer tools for humans in the era of coding agents. Its stated goal is to help engineers understand, critique, and safely manage AI-generated software through open-source code-review and architecture tools.
At a Glance
- Software engineers and staff engineers
- Teams building AI-heavy software
- Organizations using coding agents such as Claude Code and Codex
- Enterprise engineering teams concerned with code comprehension, security, and review quality
- +1 more
AI Tools by dev.fast
(1)
Whiteboard
AI Code Change Visualizer
Discussions
No discussions yet
Be the first to start a discussion about dev.fast
Latest News
Whiteboard/Review Desktop 0.1.3 released, adding Windows installers, Linux packaging improvements, oh-my-pi connectivity, and the updated data-flow/control-flow demo.
What's new in Whiteboard: September 2026 — live agent-rendered reviews, diffr structural diffs, share links, scratchpad, and agent traces.
Whiteboard/Review Desktop 0.1.2 released with Fedora download support and README guidance updates.
Whiteboard/Review Desktop 0.1.1 released with fixed packaging and Linux/macOS selection.
Products & Services
MIT-licensed open-source desktop app where humans and coding agents architect software in a common workspace. It connects to coding agents, renders interactive diagrams and semantic diffs, links visualizations to source code, and exposes agent traces and decision logs. Runs on macOS, Windows, Ubuntu, Fedora, and has Arch Linux packaging.
Alpha infrastructure that captures every agent run and makes it queryable in code, with storage tied to version control and local/VPC deployment.
Planned/early product for deciding when automated code-review tools should touch pull requests, resolving AI comments and merge conflicts, and automatically merging low-risk changes; described as OSS and BYOK.
Early product for formal verification and deterministic simulation testing, using Whiteboard infrastructure to make system specifications interpretable, expressive, and extensible.
Market Position
/dev/fast positions Whiteboard/Review as an open-source, purpose-built 'Cursor for code review' or AI-native code forge. It differentiates from conventional raw diff viewers and coding IDEs with architecture-level visualizations, AST-aware diffs, source-linked diagrams, and auditable agent traces; adjacent competitors include Cursor, GitHub/GitLab code review, and AI code-review tools.
Leadership
Founders
Sid Menon
Founder/CEO; previously a tech lead on Palantir's cloud infrastructure team, where he built secure hybrid-cloud deployment infrastructure; studied computer science at Harvard.
Milan Bhandari
Founder/CTO; cofounder of Bolto (YC S23), previously on Palantir's ML infrastructure team building tools for in-platform model deployment and evaluation; studied computer science at Harvard.
Ketan Agrawal
Founder/CSO; previously an ML engineer at Robust Intelligence, where he red-teamed AI systems, and at Snowflake, where he built data-querying agents; studied symbolic systems/computer science at Stanford.
Alex Iansiti
Founder/CPO; previously a software engineer at Pinterest and Flowcode, where he led a rearchitecture for large enterprises; studied computer science at Harvard.
Executive Team
Sid Menon
Founder/CEO
Former Palantir cloud-infrastructure tech lead; Harvard computer science.
Milan Bhandari
Founder/CTO
Cofounder of Bolto (YC S23); former Palantir ML-infrastructure engineer; Harvard computer science.
Founding Story
The founders started /dev/fast around the view that coding agents make code production faster but do not automate engineering judgment. Their initial vision is an open-source code forge for the AI era, beginning with Review/Whiteboard: a workspace where humans can understand architecture, diffs, and agent decisions before merging changes.
Business Model
Revenue Model
The current flagship Whiteboard product is open-source and self-hostable under the MIT license. The repository says a hosted product for teams is planned; no current paid pricing was published.
Target Markets
- Software engineers and staff engineers
- Teams building AI-heavy software
- Organizations using coding agents such as Claude Code and Codex
- Enterprise engineering teams concerned with code comprehension, security, and review quality
- Open-source developers and self-hosting teams
- Reviewing large or AI-generated pull requests before merge
- Understanding software architecture and system changes
- Critiquing agent-generated diagrams, diffs, and implementation decisions
- Designing new APIs and exploring implementation tradeoffs
- Auditing coding-agent traces and requirements-to-code decisions
- Formal verification and deterministic simulation testing