Build and validate a minimal Manim Studio runtime image

[alonshoa]

Problem

The current Manim Studio runtime inherits from:

manimcommunity/manim:v0.20.1

The upstream image includes a broad interactive Manim environment, including Jupyter-related and other packages that are not required for the normal Studio CLI, MCP, and render workflow.

A Docker Scout scan of the experimental 0.0.1 image reported:

• image size: approximately 1.5 GB
• packages: 817
• 2 critical vulnerabilities
• 65 high vulnerabilities
• 63 medium vulnerabilities
• 257 low vulnerabilities
• 63 unspecified vulnerabilities

Most findings originate from the upstream image, Debian packages, Jupyter components, and the broad LaTeX/font environment rather than Manim Studio code.

Goal

Investigate and validate a smaller purpose-built runtime containing only the dependencies required by Manim Studio.

Possible base:

python:3.11-slim-bookworm

Expected capabilities to preserve:

• Manim 0.20.1
• Manim Slides
• Studio CLI
• local MCP server
• Cairo/Pango text rendering
• images and video
• MathTex/LaTeX
• XeLaTeX where required
• Hebrew and RTL fonts
• named beats
• review artifacts
• PPTX export
• external-project mounts

Proposed approach

• Build a separate experimental runtime target.
• Keep the existing community-based runtime as the baseline.
• Define small capability-oriented smoke scenes.
• Compare image size, package count, Scout results, and runtime behavior.
• Consider an automated build/test minimization harness rather than manual package removal.
• Do not replace the existing runtime until the candidate passes the full capability matrix.

Acceptance criteria

• A candidate runtime builds reproducibly.
• studio doctor passes.
• Basic Manim render passes.
• MathTex render passes.
• Hebrew/RTL render passes.
• Manim Slides render passes.
• external-project verification passes.
• PPTX export passes.
• MCP startup and structured operations pass.
• Image size and package count are materially reduced.
• Critical/high Scout findings are reduced or documented.
• The final dependency list and maintenance responsibilities are documented.

Priority and relationship

This issue is non-blocking.

It must not block:

• merging Project Init
• local development
• external-project creation
• work on educational presentations

Public-runtime distribution may continue experimentally while this issue remains open.