fireworks-tech-graph is an MIT-licensed Agent Skill and Python CLI for Codex and Claude Code. It converts Chinese or English system descriptions into geometry-checked SVG, PNG, offline interactive HTML, and semantically constrained animated GIFs. It includes 12 styles, 14 UML diagram types, and structural contracts for architecture, cloud, event-flow, and observability diagrams.
Source: github.com/yizhiyanhua-ai/fireworks-tech-graph View on GitHub →It extends natural-language diagramming into a chain of structured IR, deterministic layout, geometry validation, export, and reporting. v1.2.0 adds verifiable SVG-to-GIF animation contracts for 12 styles.
Source: github.com/yizhiyanhua-ai/fireworks-tech-graph/releases/tag/v1.2.0The generator validates nodes, edges, ports, and domain fields before layout. Duplicate IDs, dangling references, invalid waypoints, and non-finite coordinates fail early.
Source: github.com/yizhiyanhua-ai/fireworks-tech-graphIt checks the canvas, labels, reserved regions, edge crossings, port fan-out, and engineering-style contracts. Synthesis does not continue after a failure.
Source: github.com/yizhiyanhua-ai/fireworks-tech-graphThe CLI produces SVG, PNG, and single-file interactive HTML. The local HTML supports zoom, themes, copying, and several bitmap downloads.
Source: github.com/yizhiyanhua-ai/fireworks-tech-graphv1.2.0 defines timelines, directions, stages, geometry, and pixel probes for 12 scenes and emits a motion.json validation report.
Source: github.com/yizhiyanhua-ai/fireworks-tech-graph/releases/tag/v1.2.0The Skill first maps a description to a style and domain structure, then schemas and IR validators normalize the input. Python generators perform layout, orthogonal routing, and port and label checks before emitting semantic SVG and a layout report. PNG output uses CairoSVG, rsvg-convert, or a browser path. Interactive export sanitizes the SVG and packages it into one HTML file. The optional animation path renders frames with Chromium and Puppeteer, composes them with FFmpeg, and validates the result through FFprobe, DOM, signature, and pixel probes plus a motion report.
Source: github.com/yizhiyanhua-ai/fireworks-tech-graph (1) · github.com/yizhiyanhua-ai/fireworks-tech-graph/blob/v1.2.0/CHANGELOG.md (2)Python 3.9+CairoSVG or rsvg-convert for PNGThe optional GIF path requires FFmpe…The skills CLI installation path req…It fits architecture review diagrams, cloud deployment diagrams, event flows, reliability troubleshooting diagrams, UML, and technical-document illustrations, with reports available for regression checks. Mermaid is lighter for simple relationships embedded in Markdown; draw.io is better for freeform dragging and manual layout.
Source: github.com/yizhiyanhua-ai/fireworks-tech-graphv1.2.0 is the latest stable release, version date 2026-07-17. It adds 115 frames of default semantic animation, motion report, Dom immutability protection, running probes, FFprobe checks, and 852 Chromium compatibility comparisons for 12 styles; static workflows do not require animation dependencies.
Source: github.com/yizhiyanhua-ai/fireworks-tech-graph/releases/tag/v1.2.0status: docs-only | checked_at: 2026-09-23 | environment: Check the official README, package.json, v1.2.0 Release/changelog, and shallow clone the main check directory; skill installer is not running, CairoSVG/Puppeteer/FFmpeg is not installed, no image is generated | limitations: No local installation or runtime verification was performed. Unverified Installation, Dependency Parsing, Any Layout, SVG/PNG/HTML/GIF Output, Geometry Check, 852 Compatibility Comparisons, Pixel Threshold or Visual Quality | observed_output: The latest stable release is v1.2.0, dated 2026-07-17. package.json remains at 1.2.0, while the official documentation says the npm channel is still at 1.0.4.
Source: github.com/yizhiyanhua-ai/fireworks-tech-graph/releases/tag/v1.2.0 (1) · github.com/yizhiyanhua-ai/fireworks-tech-graph/blob/main/package.json (2)Compared to Mermaid, fireworks-tech-graph generates finished diagrams with fixed styles, geometric checks, exported reports, and animated contracts from natural language and structured JSON; Mermaid uses a concise DSL that is natively suitable for Markdown, document stations, and common flow/UML diagrams. Select Mermaid for readable diff, lightweight embedding, and fireworks-tech-graph for 1920px technical illustrations, offline interactive files, or tested animations.
Source: github.com/yizhiyanhua-ai/fireworks-tech-graph (1) · github.com/mermaid-js/mermaid (2)This Skill has clear value when repeatable, high-resolution technical illustrations and strict export checks are required. Pin v1.2.0, run the fixture through static SVG first, and add PNG or GIF dependencies only as needed. Treat automated checks as review aids; people still need to confirm the content and readability.
Source: github.com/yizhiyanhua-ai/fireworks-tech-graphThe current stable Release is v1.2.0. The official documentation says the npm registry remains at 1.0.4, so use the v1.2.0 Release archive or the nested Skill path at a pinned tag and verify SHA256SUMS.
No. Static SVG and offline HTML use the Python toolchain. PNG requires CairoSVG or rsvg-convert. Chromium, Puppeteer, and FFmpeg/FFprobe are only required for the optional GIF animation path.
No. Structural and geometry validation can catch dangling references, overflow, crossings, and related faults, but visual judgment still requires reading the image.
No. The sponsor promotion has no technical dependency on the local Skill, Python CLI, CairoSVG, Puppeteer, or FFmpeg. Treat it as a sponsorship link rather than an installation requirement.