Skip to main content
Glama
SalZaki

doodleworks-mcp

by SalZaki
OverviewSchemaSourceRelated ServersScoreDiscussionsAdmin
TypeScript
Hybrid

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
PORTNoPort for the dev Streamable HTTP server3001
GEMINI_API_KEYNoGemini API key for Gemini 3 Pro Image
GOOGLE_API_KEYNoGoogle API key (alias for GEMINI_API_KEY)
OPENAI_API_KEYNoOpenAI API key for GPT Image (default provider)
DOODLEWORKS_QUALITYNoImage quality: low | medium | high | autolow
DOODLEWORKS_STYLE_REFNoDefault style-reference id/path
DOODLEWORKS_CONCURRENCYNoNumber of illustrations rendered in parallel3
DOODLEWORKS_REQUEST_TIMEOUT_MSNoPer-image request timeout in milliseconds120000

Instructions

Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.

This server publishes no instructions, or was last inspected before Glama recorded them.

Capabilities

Features and capabilities supported by this server

Protocol revision2025-11-25

CapabilityDetails
tools
{
  "listChanged": true
}
prompts
{
  "listChanged": true
}
resources
{
  "listChanged": true
}

Tools

Functions exposed to the LLM to take actions

NameDescription
create_illustrationsA

Render a set of single-idea, hand-drawn Tinku illustrations (each 16:9; an optional 21:9 hero) and open an interactive viewer. Supply one entry per illustration; each prompt is a contraption composition + Tinku's action + its exact English annotations — the server applies the Tinku character, the conceptual engine, and the house style. Use this when the user wants hand-drawn illustrations for an article, post, or explainer.

get_illustrationA

Return the rendered image for one illustration of a previously created set. The viewer calls this for each illustration after create_illustrations so a single tool result never carries the whole set.

regenerate_illustrationB

Re-render a single illustration with the same (or a tweaked) prompt and aspect. Image models are non-deterministic, so this yields a fresh take. Returns the new image in _meta for the viewer.

Prompts

Interactive templates invoked by user choice

NameDescription
plan_illustrationsTurn a topic or source into a create_illustrations-ready set of single-idea Tinku illustrations: extract cognitive anchors, reinvent each as a contraption Tinku operates, English annotations only, finishing with one create_illustrations call.

Resources

Contextual data attached and managed by the client

NameDescription
ui://doodleworks-mcp/viewer.html
visual-dnaThe Tinku house style: pure-white background, clean even black line, sparse red/orange/blue handwritten English annotations.
concept-engineCognitive anchor → low-tech contraption Tinku operates; the erase-test; no literal topic pictures.
composition-patternsSecondary layout aids (process, cycle, stack, taxonomy, matrix, timeline, decision, data-shape, summary) for arranging a contraption — never a literal chart.
prompt-patternsHow to write an illustration prompt: contraption + Tinku's action + the Required-text block (red/orange/blue, terse).
intakeQuestions to scope a set: topic/source, audience, spine, count, the cognitive anchors to illustrate.
output-qualityThe quality bar and a pre-send checklist (erase-test, pure-white, sparse annotations, 16:9, no flowchart).
style-referencesThe drawing-style reference ids available to pass as styleReference (style calibration only — not composition templates, and not the source of the Tinku character).