Skip to the content
Know your tech before you buy it. Makers: get listed from $49/mo

Lucid Train

lucidtrain.com

  • Free plan
  • Windows
  • macOS
  • Linux
7.6 Club score Visit lucidtrain.com

What Lucid Train is

Lucid Train is a desktop coding harness for people who want to map a repository and use its architecture diagram to guide a coding agent. It draws diagrams from a codebase and passes them to the agent as specifications. The agent emits a semantic graph, and the ELK layout engine positions diagram elements and routes their connections. Diagrams are Excalidraw scenes that can be exported as SVG or PNG, or converted into Markdown specifications; Mermaid support and diagram import are also listed. Proposed edits appear as diffs, and commands are sorted into risk tiers, with catastrophic commands refused. Lucid Train supports local Ollama models, OpenRouter, other OpenAI-compatible endpoints and hosted cloud models. Repository files are processed locally, while prompts and code go to the selected model provider or stay local with local models. The maker says it collects no telemetry, analytics, prompts, code or usage data from local or bring-your-own-key sessions. API keys stay in the local configuration directory or OS keychain. Connectors include GitHub, Vercel, Railway and Grafana; MCP connections use stdio only, not remote HTTP or SSE.

Who it suits

Lucid Train suits developers who want to understand a repository's architecture before directing a coding agent. It is aimed at people who prefer a systems-design-first workflow and can use local models or their own provider keys.

Good to know, and what to watch

  • Builds architecture diagrams from repository code.
  • Exports diagrams to SVG, PNG and Markdown.
  • Shows proposed edits as diffs.
  • Supports local Ollama and hosted model providers.
  • API keys remain in local storage or the OS keychain.
  • MCP connections support stdio only.
  • Desktop license covers up to three machines.
  • Desktop sign-in is required for licensing.

CLI

Free

Lucid Train

$2.50/mo

Free plan · 7-day trial

Getting started

  1. Open the Lucid Train website
  2. Install the CLI or desktop app for Linux, macOS, or Windows
  3. Use the free CLI without an account, or start the desktop trial
  4. Connect a local Ollama model, OpenRouter, another OpenAI-compatible endpoint, or a hosted model
  5. Open a repository and generate its architecture diagram
  6. Add listed connectors or install stdio MCP servers as needed

Limits to check

The CLI plan is free forever but provides the listed CLI agent capabilities. The Lucid Train desktop plan costs $2.50 a month when billed annually, includes a seven-day trial with no card up front, and each desktop license covers up to three machines.

The Club's review

Lucid Train is a desktop and CLI tool that turns repository architecture into a guide for coding agents. It is best suited to developers who want system design to shape agent-assisted code changes. Choose it for that focused workflow; look elsewhere for a general-purpose diagram canvas or remote MCP connections.

Overview

Lucid Train links repository mapping to implementation: it draws an architecture diagram from a codebase, then passes that diagram to a coding agent as a specification. That connection is its main distinction. It is a better fit for planning work against an existing codebase than for teams seeking a standalone diagramming workspace.

Proposed edits are shown as diffs, and commands are assigned risk tiers, with catastrophic commands refused. These controls make changes easier to review and block the highest-risk commands, but they do not replace inspecting an agent's proposed edits.

Key features

  • Architecture mapping: An agent emits a semantic graph, and the ELK layout engine calculates diagram positions and orthogonal edge routes. The result is a structured view that can guide coding-agent work, rather than a diagram that ends at documentation.
  • Diagram formats: Diagrams are Excalidraw scenes, exportable as SVG or PNG, and can also become Markdown specifications. Mermaid support and diagram import add flexibility for users bringing other diagram workflows into the tool.
  • Model options: Lucid Train supports local Ollama models, OpenRouter, other OpenAI-compatible endpoints and hosted cloud models. Local processing means repository files stay on the machine, while prompts and code go to the selected provider—or nowhere when using a local model.
  • Connections: The Connectors tab lists GitHub, Vercel, Railway and Grafana, and can install arbitrary stdio MCP servers. The stdio-only constraint rules out direct connections to remote HTTP and SSE servers, which may be a blocker for teams whose MCP services use those transports.
  • Privacy and credentials: The maker says it collects no telemetry, analytics, prompts, code or usage data from local or bring-your-own-key sessions. API keys are stored in the Lucid Train configuration directory or OS keychain and are not sent to the maker. These practices suit users who want control over model choice and credential storage; cloud-model prompts and code still go to the chosen provider.

Pricing

The free CLI plan costs nothing, billed Free · forever. It includes the full agent, local models or your own API keys, is open source and requires no account. It is the sensible starting point for developers who need the agent workflow but do not need a native desktop app.

The Lucid Train plan costs $2.50 a month, billed billed $30 a year. It includes everything in the CLI in a native app, plus Design, Diagram, Audit and Dashboards. A 7-day free trial is offered with no card up front. It fits users who want those desktop tools alongside the agent workflow; the desktop license covers up to three machines and requires desktop sign-in for licensing.

Platforms

Lucid Train is available for Linux, macOS and Windows as desktop software, with a CLI option. The desktop license's three-machine limit matters for users who switch among more devices.

Who it's for

Lucid Train is aimed at systems-design-first developers who want to map a repository and use that map to guide a coding agent. The free CLI suits people comfortable with local models or their own API keys; the paid plan is for those who also want the native app's design, diagram, audit and dashboard tools. It is less suitable when remote HTTP or SSE MCP connections are essential.

Pros and cons

Pros

  • Architecture informs agent work: The generated diagram is a specification for coding, making repository mapping part of the implementation workflow.
  • Useful output choices: SVG, PNG and Markdown exports serve visual sharing and text-based handoffs, while the underlying diagrams are Excalidraw scenes.
  • Broad model choice with local options: Local Ollama and bring-your-own-key setups let users choose where inference happens and how credentials are stored.
  • A capable no-cost entry point: The free CLI includes the full agent, supports local models or personal API keys, and needs no account.

Cons

  • MCP transport is limited: Only stdio connections work directly, so remote HTTP and SSE servers are out of reach.
  • Desktop licensing has a device cap: One license covers up to three machines and requires sign-in, a constraint for users with larger device fleets.
  • Cloud-model use sends prompts and code to the provider: Users who need those inputs to stay local must choose a local model.

Alternatives

  • Creately is a freemium alternative with web and self-hosted options, unlimited canvases and collaborators on its free plan, but caps each canvas at 45 items and one fold. Choose it for collaborative canvases across platforms rather than Lucid Train's repository-to-agent workflow.
  • ClearWork is a paid web option with read-only access for up to 10 users included on every core plan. Consider it when web access and a read-only audience matter more than desktop repository mapping.
  • Capci offers a free plan with 50 lifetime generations shared across posts, diagrams and animations, one brand, and API and MCP inclusion. It may suit users looking for capped, multi-format generation rather than a codebase-focused harness.
  • Mermaid Chart has a free tier capped at three diagrams, with limited AI, one team and one project; its Plus plan includes unlimited diagrams and 300 AI credits. Choose it when a web or extension-based Mermaid diagram workflow is the priority.
  • Splotch offers one free project with 10 diagrams, two-way sync, manual and AI editing, sharing and export, plus Google Docs import and a plugin. It is a more relevant option for that synced, document-connected diagram workflow.
  • Archflow has a free web plan with one project, three diagrams per project, user stories with a Kanban board and a basic UML editor. Choose it for that project-and-user-story workflow instead of repository diagrams handed to coding agents.
  • patchley has a free plan with five diagram generations for flowcharts or BPMN 2.0, plus limited audio and text inputs. It is an option for generating those diagram types from short inputs.
  • OpenCharts offers a free tier with 500K AI credits, one of each project type, one Code Canvas project, 10 GB of Drive storage and all export formats. Consider it if those project and storage allowances are a better fit than Lucid Train's focused architecture-to-agent workflow.

For a broader directory of tools, browse AI Diagram Generators or AI Wireframe Generators.

Verdict

Lucid Train is a strong fit for developers who want repository architecture to guide coding-agent work, especially those who value a free CLI and a choice of local or hosted models. Its diagram-to-specification workflow is the reason to choose it. Look elsewhere if you need direct remote HTTP or SSE MCP connections, or if the three-machine desktop license is too restrictive.

The verdict

Lucid Train connects repository architecture mapping with a coding-agent workflow and supports several model options. Its stdio-only MCP connections and three-machine desktop license limit are worth weighing before adopting it.

Questions people ask

Is Lucid Train free?

The CLI plan is free forever, open source, and requires no account. The Lucid Train desktop plan is $2.50 a month when billed annually and has a seven-day free trial.

Which platforms does it support?

The desktop software supports Linux, macOS, and Windows.

Which model providers can it use?

It supports local Ollama models, OpenRouter, other OpenAI-compatible endpoints, and hosted cloud models.

How are repository files and credentials handled?

Repository files are processed locally. API keys are stored in the Lucid Train configuration directory or operating-system keychain and are not sent to the maker.

Can it connect to remote MCP servers?

No. MCP connections use stdio only, so remote HTTP and SSE servers cannot connect directly.

What does the paid desktop plan add?

The native app includes everything in the CLI plus Design, Diagram, Audit, and Dashboards, with local models or your own API keys.

What the maker publishes

Agent workflow
Edits are presented as diffs, and commands use risk tiers with catastrophic commands refused. source
Architecture diagrams
The agent emits a semantic graph and the ELK layout engine computes diagram positions and orthogonal edge routing. source
Credentials
API keys are stored locally in the Lucid Train configuration directory or OS keychain and are not sent to the maker. source
Data handling
The software processes repository files locally, and prompts and code go directly to the chosen model provider or nowhere when local models are used. source
Desktop license limit
A desktop license covers up to three machines, and desktop sign-in is required for licensing. source
Diagram exports
Diagrams are Excalidraw scenes that export to SVG and PNG and can also be turned into a Markdown specification. source
Integrations
The Connectors tab lists GitHub, Vercel, Railway and Grafana and can install arbitrary stdio MCP servers. source
Intended users
The product is described as a systems-design-first coding harness for people who want to map a repository and use that diagram to guide a coding agent. source
MCP limitation
MCP connections use stdio only; remote HTTP and SSE servers cannot be connected directly. source
Model support
Lucid Train supports local Ollama models, OpenRouter and other OpenAI-compatible endpoints, plus hosted cloud models. source
Product
Lucid Train draws architecture diagrams from a codebase and hands the diagram to a coding agent as a specification. source
Support
The maker lists [email protected] for questions about its terms and privacy policy. source
Telemetry
The maker says it does not collect telemetry, analytics, prompts, code or usage data from local or bring-your-own-key sessions. source

Checked against the maker's pages on 5 Oct 2026.