Lucid Train compared with Windsurf.
Both lean agentic. One of them has a canvas and runs on any model you like.
Lucid Train generates architecture diagrams from your codebase, uses them as implementation specifications, and runs against any OpenAI-compatible model including fully offline local ones. It can also drive Claude Code, Codex, Cursor or OpenCode as its execution engine. Windsurf is an agentic AI IDE where the agent works inside the editor on its own hosted models.
A visual design surface
Windsurf's agent is good at working through a task in code and has no canvas for deciding what the system should be. Lucid Train generates the architecture from the repository, lets you edit it as a graph with the layout re-solving, and converts it into a specification that starts a coding turn.
Any model, offline, and other agents as engines
Windsurf is a hosted product on its models. Lucid Train runs against any OpenAI-compatible endpoint including local Ollama with nothing leaving the machine, assigns different models to different roles, and can drive Claude Code, Codex, Cursor or OpenCode as the engine.
Harness depth
Sub-agents executing in isolated git worktrees with the diff applied back, concurrent read-only explorers, automatic context compaction, persistent offline memory, Claude Code-compatible skills, and a publish-state guard that blocks destructive commands once an acceptance check has passed.
More than code in one window
The desktop app also carries a dependency audit, a Grafana, Prometheus and Loki surface, and connectors, all driven by the same agent core, with multiple project workspaces running their own sidecars concurrently and true multi-window support.
What Windsurf is good at
It is a full IDE with an agent that maintains context across a session and executes multi-step work where the code lives. For the write-run-fix loop, being inside the editor is a real advantage.
Side by side
| Lucid Train | Windsurf | |
|---|---|---|
| Architecture diagrams from code | Yes | No |
| Diagram as an implementation specification | Yes | No |
| Runs other agents as engines | Claude Code, Codex, Cursor, OpenCode | No |
| Bring your own model | Any OpenAI-compatible endpoint | Limited |
| Fully offline | Yes, with local models | No |
| Sub-agents in isolated git worktrees | Yes | No |
| Observability surface | Grafana, Prometheus, Loki | No |
| Dependency audit | Yes | No |
| Multi-window | Yes | Editor windows |
| Full editor | No | Yes |
| Price | CLI free and open source; desktop $3/month, 7-day trial, no card | Free tier, then subscription |
Questions
Is Lucid Train an IDE?
No. It has a coding agent producing reviewable diffs, not a text editor with a debugger and an extension ecosystem.
Which for greenfield work?
Design in Lucid Train first, because architecture decisions are cheapest to change before any code exists, then implement in whichever editor you prefer.
Can it run fully offline?
Yes, against local Ollama models with no API key and no telemetry.
Affiliation
Lucid Train is an independent product and is not affiliated with, endorsed by, or connected to the product described on this page. All names and trademarks belong to their respective owners. Pricing and features were checked on 19 August 2026 and change without notice; check the vendor's own site before deciding.
Related
- Lucid Train vs AiderThe closest match on model philosophy. The difference is everything above the coding loop.
- Lucid Train vs IcePanelIcePanel models what you intend. Lucid Train reads what you built.
- Lucid Train vs CodeVizSame starting point, generated from code. Different answers to what the diagram is for.
- Lucid Train vs GitDiagramGitDiagram needs no install to look at a public repo. Lucid Train is for the one you work in every day.
Last updated