A
licenseB
qualityA
maintenanceAn MCP server that generates beautiful Excalidraw architecture diagrams with perfect auto-layout, stateful editing, and architecture-aware component styling.
26
313 PyPI
151
MIT
sketchboard-excalidraw-mcp
by sandraschi
Currently, the server is mostly a scaffold: you can query its status and see registered tools, but the rendering/conversion/batch tools return not_implemented.
status: reports the server identity, version, and which capability groups are wired up.render_svg: registered to render Excalidraw element JSON to SVG, but currently returnsnot_implemented.mermaid_to_excalidraw: registered to convert Mermaid source to Excalidraw elements, but currently returnsnot_implemented.batch_render: registered to batch-render README diagrams across a list of repos, but currently returnsnot_implemented.
Allows creation and rendering of Excalidraw diagrams, including interactive canvas authorship and headless generation of SVG/PNG from Excalidraw element JSON.
Converts Mermaid diagrams to Excalidraw format for editing and rendering within the server (planned).
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@sketchboard-excalidraw-mcpdraw a simple network topology diagram"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
# sketchboard-excalidraw-mcp
Excalidraw canvas authorship + headless-render MCP server. Slots next to inkscape-mcp and gimp-mcp as the vector/raster/diagram trio in the fleet.
Status: day 2. Server shell, package layout, and fleet standardization
are in place. Headless rendering (render_svg/render_png) is real and
tested. Canvas bridging, Mermaid conversion, and batch rendering are
registered as MCP tools but return not_implemented - see docs/DESIGN.md
for what's real vs planned.
What this is
A merge of two upstream approaches into one fleet-standard server:
Canvas workbench (planned base: fork of lesleslie/excalidraw-mcp, itself built on yctimlin/mcp_excalidraw): Python FastMCP server + TypeScript canvas server + React frontend for live, interactive diagram authorship. Not merged in yet - both upstreams are cloned for reference/cherry-picking under
D:\Dev\repos\external\, but no canvas-bridge code has been ported into this repo. See the fork-drift open question indocs/DESIGN.mdbefore that lands.Headless rendering (real, implemented): ported from bassimeledath/excalidraw-render-mcp's TypeScript/agent-browser renderer into Python/Playwright, rather than vendoring a second Node stack. A warm Chromium instance renders Excalidraw element JSON to SVG/PNG via
exportToSvg, reused across calls in the same process. Known gap: still loads@excalidraw/excalidrawfrom esm.sh at runtime rather than a locally vendored bundle, so it needs network access on first render and isn't version-pinned yet.
The two modes are deliberately decoupled: headless render tools work without the canvas server running (agent batch mode), while the canvas/Tauri app is for interactive human use. Both talk to the same backend; the desktop app is not a sidecar wrapping the Python server.
Related MCP server: excalidraw-export-mcp
Fork sources on this machine
Both upstream forks were cloned for reference/cherry-picking, not vendored wholesale:
D:\Dev\repos\external\excalidraw-mcp-lesleslie\D:\Dev\repos\external\excalidraw-render-bassimeledath\
Planned extras
Mermaid-to-Excalidraw conversion, wrapping the official
@excalidraw/mermaid-to-excalidrawnpm package rather than reimplementing Mermaid parsing..excalidrawlibsupport for reusable fleet-diagram components (server boxes, transport arrows).Batch-render command for the repo-README diagram pipeline across the fleet (~200 repos).
Ports
Port | Service |
11107 | Backend (FastMCP HTTP |
11108 | Frontend (Vite React canvas) |
Registered in mcp-central-docs/operations/WEBAPP_PORTS.md.
License and attribution
MIT, same as both upstream sources. Credit chain: this project builds on
lesleslie/excalidraw-mcp (which credits yctimlin/mcp_excalidraw) and borrows
the headless-render architecture from bassimeledath/excalidraw-render-mcp.
Font licensing (Excalidraw's default hand-drawn fonts ship under SIL OFL, not
MIT) is tracked separately in NOTICE.md once the JS bundle is vendored.
Development
uv sync
uv run pytest
just release-drySee docs/DESIGN.md for the day 1/2/3 build plan and the concrete technical
decisions behind this scaffold.
Available Tools
4 toolsbatch_renderA
Batch-render README diagrams across a list of fleet repos.
NOT IMPLEMENTED YET. Day 2/3 work: browser-context pool over the one warm Chromium instance (not one page reused sequentially), so a 200-repo run doesn't serialize behind a single tab.
Return Format
{"success": False, "error": str, "error_type": "not_implemented", "suggestions": list[str]}
ParametersJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| output_dir | Yes | Directory to write rendered diagrams into. | |
| repo_list_path | Yes | Path to a text file listing one repo per line. |
Output Schema
ParametersJSON Schema
| Name | Required | Description |
|---|---|---|
No output parameters | ||
TDQS
A3.8/5.0
Behavior4/5
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries full responsibility and clearly discloses that the tool is not implemented, will always return an error with a fixed format, and includes the exact error structure. It also explains planned architecture, which is extra context beyond the minimal stub behavior.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Conciseness3/5
Is the description appropriately sized, front-loaded, and free of redundancy?
The first sentence and return format are concise, but the 'Day 2/3 work' paragraph contains speculative implementation details that are not essential for an agent using the tool. This adds mild bloat to an otherwise well-structured stub description.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Completeness4/5
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a not-implemented stub, the description fully informs the agent that the tool will fail and what response to expect. It even includes a future roadmap. This is complete given the tool's current state, though if the tool were implemented, more behavioral details would be needed.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Parameters3/5
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema covers both parameters completely with descriptive text for each (output_dir and repo_list_path), so baseline is 3. The description adds no additional parameter semantics beyond the schema's already clear definitions.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Purpose5/5
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's function with a specific verb ('Batch-render'), resource ('README diagrams'), and scope ('across a list of fleet repos'). This distinguishes it from sibling tools like render_svg, which presumably handles single renders.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Usage Guidelines3/5
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description explicitly warns 'NOT IMPLEMENTED YET,' which is crucial guidance against using the tool. However, it does not explicitly mention alternatives or provide when-to-use vs. when-not-to-use guidance beyond this warning.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
mermaid_to_excalidrawA
Convert a Mermaid diagram to Excalidraw elements and optionally render it.
NOT IMPLEMENTED YET. Planned to wrap the official @excalidraw/mermaid-to-excalidraw npm package via a small Node/Bun shim rather than reimplementing Mermaid parsing.
Return Format
{"success": False, "error": str, "error_type": "not_implemented", "suggestions": list[str]}
ParametersJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| output_path | Yes | Path to write the converted/rendered output. | |
| mermaid_source | Yes | Mermaid diagram source text. |
Output Schema
ParametersJSON Schema
| Name | Required | Description |
|---|---|---|
No output parameters | ||
TDQS
A4.3/5.0
Behavior5/5
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are present, so the description bears full responsibility. It discloses the stub behavior and exactly what return format to expect (success: False, error_type: 'not_implemented', suggestions), plus explains the planned implementation. No contradictions.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Conciseness4/5
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is front-loaded with the purpose and the critical 'NOT IMPLEMENTED' notice. The implementation-plan paragraph adds context but is arguably unnecessary; still, the overall structure is organized and not overly verbose.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Completeness4/5
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For an unimplemented tool, the description fully covers what an agent needs to know: the tool will fail gracefully with a specific error format. The optional rendering mention is vague since there is no render parameter, but that is minor given the not-implemented status.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Parameters3/5
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema already provides clear descriptions for both parameters (output_path and mermaid_source) with 100% coverage. The description does not add additional parameter-level meaning, so the baseline of 3 applies.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Purpose5/5
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the intended function: converting Mermaid diagrams to Excalidraw elements and optionally rendering. It is specific and distinguishes from sibling tools (status, render_svg, batch_render).
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Usage Guidelines4/5
Does the description explain when to use this tool, when not to, or what alternatives exist?
Explicitly states 'NOT IMPLEMENTED YET', which is a clear instruction that the tool should not be used for actual conversions. While it doesn't name alternatives, it provides strong contextual guidance on when to avoid the tool.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
render_svgA
Render Excalidraw element JSON to SVG via headless Chromium (Playwright).
NOT IMPLEMENTED YET. Day 2 work: port bassimeledath's warm-Chromium exportToSvg approach into sketchboard_excalidraw_mcp/render.py using a locally vendored excalidraw JS bundle (no CDN dependency at runtime).
Return Format
{"success": False, "error": str, "error_type": "not_implemented", "suggestions": list[str]}
ParametersJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| output_path | Yes | Path to write the output SVG file. | |
| elements_json | Yes | Excalidraw elements as a JSON array string. |
Output Schema
ParametersJSON Schema
| Name | Required | Description |
|---|---|---|
No output parameters | ||
TDQS
A3.9/5.0
Behavior5/5
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. It fully discloses that the tool is not implemented, that it will always return a failure object with 'success': False, and specifies the error_type and suggestions fields. This is complete transparency for a stub tool, leaving no ambiguity about its runtime behavior.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Conciseness2/5
Is the description appropriately sized, front-loaded, and free of redundancy?
The description contains unnecessary implementation details about 'Day 2 work', 'bassimeledath's warm-Chromium approach', and 'locally vendored excalidraw JS bundle' that do not help an agent select or invoke the tool. The return format is clearly structured, but the clutter makes it less concise than it should be.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Completeness4/5
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a two-parameter tool with an output schema (return format), the description adequately covers the essential context: the not-implemented status, the underlying technology, and the exact failure return format. The main gap is the absence of any note about what the tool would do if it were implemented, but since it is a stub, the current information is sufficient.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Parameters3/5
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The schema already provides descriptions for both required parameters (elements_json and output_path) with 100% coverage. The tool description adds no additional meaning or nuance beyond what the schema states, so the baseline score of 3 is appropriate.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Purpose5/5
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool 'Render Excalidraw element JSON to SVG via headless Chromium (Playwright)', providing a specific verb, resource, and method. This distinguishes it from sibling render tools like batch_render by indicating it handles a single conversion, and the 'NOT IMPLEMENTED YET' status clarifies its current non-functional state without obscuring the intended purpose.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Usage Guidelines3/5
Does the description explain when to use this tool, when not to, or what alternatives exist?
The 'NOT IMPLEMENTED YET' warning implicitly signals the agent not to use this tool until it is ready, which is a form of usage guidance. However, it does not explicitly state when it should be used, nor does it point to alternatives such as batch_render for current rendering needs, leaving the when-to-use guidance incomplete.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
statusA
Report server identity and which capability groups are wired up yet.
This is the one real tool in the day-1 scaffold - everything else is registered but returns an explicit not_implemented result.
Return Format
{"server": str, "version": str, "capabilities": dict[str, str]}
Examples
status()
ParametersJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
Output Schema
ParametersJSON Schema
| Name | Required | Description |
|---|---|---|
No output parameters | ||
TDQS
A4.5/5.0
Behavior4/5
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries the burden of explaining behavior. It discloses the return format, the fact that it's the sole functional tool, and that other tools are stubs. It does not mention side effects, but for a status read tool that is acceptable.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Conciseness5/5
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is extremely concise: two lead sentences, a return format block, and an example. It is front-loaded with purpose and includes only essential context.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Completeness5/5
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a zero-parameter status tool, the description fully covers what it does, what it returns, and how it fits into the larger toolset. The provided return format and example make it complete for an agent.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Parameters4/5
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The tool has zero parameters, so schema coverage is 100% and the description correctly adds no parameter details. The baseline score of 4 applies because no parameter ambiguity exists.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Purpose5/5
Does the description clearly state what the tool does and how it differs from similar tools?
The description opens with a specific verb 'Report' and a clear resource: 'server identity and which capability groups are wired up yet.' This unambiguously states the tool's function and distinguishes it from the sibling rendering tools.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Usage Guidelines4/5
Does the description explain when to use this tool, when not to, or what alternatives exist?
It provides the context that this is the only functional tool in the scaffold, implying that all others return not_implemented. This tells the agent when to use it (when wanting real functionality) but does not explicitly name alternatives or give a fuller when-not-to-use scenario.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
4 tool updates
v0.1.0- First observed
batch_render - First observed
mermaid_to_excalidraw - First observed
render_svg - First observed
status
TDQS
A3.9/5.0
Scored across 4 tools
Disambiguation5/5
The four tools have clearly distinct purposes: status for server info, render_svg for single-file rendering, mermaid_to_excalidraw for format conversion, and batch_render for multi-repo processing. Even though three are unimplemented, their intended roles are unambiguous and non-overlapping.
Naming Consistency4/5
All names use snake_case and mostly follow a verb_noun pattern (render_svg, batch_render), with mermaid_to_excalidraw being a noun_to_noun exception and status being a simple noun. The pattern is mostly consistent and readable, only minor deviations.
Tool Count5/5
Four tools is a well-scoped number for this server's purpose, covering status, single render, conversion, and batch processing. The count is neither too thin nor too heavy, and each tool has a clear intended role.
Completeness1/5
Three of four tools are marked NOT IMPLEMENTED YET and return explicit not_implemented errors. This creates severe dead ends for any agent attempting real diagram conversion or rendering; the only functional tool is status, leaving the server's stated purpose almost entirely unfulfilled.
Maintenance
ActivitySlowing
ResponsivenessNo issues
Related MCP Connectors
Collaborative whiteboard MCP server — create objects, connectors, C4 diagrams, and manage boards
Render, verify, describe, and safely edit Mermaid diagrams through MCP.
MCP server for progressive tool usage at any scale (see https://klavis.ai)
Related MCP Servers
- AlicenseAqualityDmaintenanceAn MCP server that enables exporting Excalidraw .excalidraw files to PNG or SVG format, with support for batch export and metadata retrieval.3MIT
- FlicenseNot gradedqualityCmaintenanceMCP server that streams hand-drawn Excalidraw diagrams with smooth viewport camera control and interactive fullscreen editing.-
- AlicenseAqualityDmaintenanceMCP server that converts Mermaid diagrams to styled Excalidraw files, saving them directly to disk without token overhead.2MIT