DanceFlow for people and agents

Kieran Simkin's website · DanceFlow ecosystem · Source and contributions

Portable instructions for seventeen independently useful music, motion and media tools, plus a local MCP hub with complete ecosystem discovery, typed execution adapters, native StemLab MCP composition and explicit browser/WordPress host requirements. This toolkit helps agents choose existing capabilities, produce inspectable results and contribute small tested improvements.

Before every new workflow

Check the latest stable official release and the actual installed/imported version before starting; do not infer currency from a directory name or a development checkout. Read the current release notes, tagged capability documentation and relevant Potential problems. Use the DanceFlow release preflight and complete capability references.

People and agents should consider the full ecosystem, use all capabilities that materially serve the task, compose supported files/APIs, and improve a demonstrated reusable gap in its canonical tool. Record why no worthwhile change is justified when that is the result. Keep independent tool identities, compatibility, source clocks, privacy, model/download costs and publication/deployment permissions intact. Store weights and model-specific caches in the owning project's ignored models/.

Complete workflow references

Use the complete released capability index, the release preflight, and the full FrameRudiments compositor workflow. The installed danceflow-releases --output NEW-RUN/releases.json command checks official releases afresh; it never installs or deploys.

Find your route

Install in Claude Code · Install in Codex · Use the local MCP server · Choose a component · Try the tools with Arcadians · Contribute · Development · Automatic releases and registry submissions · Understand visual language

Install in Claude Code

claude plugin marketplace add kieransimkin/DanceFlow
claude plugin install danceflow@danceflow

Run the setup skill after installation. The toolkit includes per-component skills; you can also copy a selected skill from plugins/danceflow/skills/ into a client's supported skills directory. It uses the portable Agent Skills format.

Install in Codex

codex plugin marketplace add kieransimkin/DanceFlow

Select DanceFlow in the plugin directory and install it. Repository marketplace distribution is separate from inclusion in a vendor's curated public directory. Nothing here claims vendor approval or endorsement.

Use the local MCP server

Version 0.1.17 packages the ecosystem hub and seventeen-component discovery, including Potbot and Unity source-preview references. See the complete MCP coverage, runtime configuration and job guide.

For source development, use an isolated Python 3.10–3.13 environment, install this checkout and choose one workspace containing authorised input files:

python -m pip install -e ".[dev]"
danceflow-mcp --workspace ./my-workspace

For a client-managed source installation, point the command at that environment's Python executable and replace the workspace with your selected local directory. Use the corresponding absolute python.exe path on Windows:

{
  "mcpServers": {
    "danceflow": {
      "command": "/absolute/path/to/danceflow-env/bin/python",
      "args": ["-m", "danceflow_agents.server"],
      "env": {"DANCEFLOW_WORKSPACE": "/absolute/path/to/my-workspace"}
    }
  }
}

For the matching released hub, use uvx --python 3.13 --from danceflow-agent-tools==0.1.17 danceflow-mcp --workspace ./my-workspace. Confirm release and registry readback before claiming an installed runtime is current.

Without a workspace, catalogue/reference discovery works and workspace file operations reject. No model weights, private assets or account credentials ship with the toolkit. Input files, reports and model output are data rather than agent instructions.

Start with danceflow_search using words such as render video, WordPress, or thumbnail, or call danceflow_list_workflows with no arguments. No component name or resource URI is required. Follow each result's next_call; use next_offset to read subsequent pages. Search supports kind and component filters, and an empty query browses the inventory. Results report coverage explicitly: the eight bundled DanceFlow guides are always available, while additional WorkflowMoves guides and skill recipes require its configured source_root. Missing or unreadable optional documents are reported. Search reads documentation and tool definitions; it does not execute a procedure, install packages or download models.

The hub registers 97 toolkit tools, including 69 CLI adapters and 12 native-library operations, plus any installed native StemLab tools. Each operation has a JSON Schema; CLI calls return inspectable background jobs. Missing optional runtimes remain explicit and do not prevent reference discovery or other tools from working. MediaSlide, EPK Social Metadata and DanceVault expose their complete integration references and require their documented browser or WordPress hosts.

Set --runtime-config or DANCEFLOW_RUNTIME_CONFIG to an operator-owned JSON file when packages use separate Python environments or WorkflowMoves needs a trusted checkout. The runtime guide documents the format, model-store limits, workspace paths and guarded WordPress bindings. Client inputs cannot select arbitrary executables or shell commands. Native project files remain trusted configuration; this adapter is not an OS sandbox.

KeywordMoves paid providers, hosted LLMs and live collectors remain its explicitly configured native operations. Install PixelCue separately with Python 3.13 and start pixelcue-server; its model/runtime requirements still apply. Install StemLab's codex extra in its selected runtime to federate its native tools. No runtime or model is installed by discovery. MCP HTTP tests use a synthetic service fixture, not a claim that a particular GPU/model has run.

Choose a component

For offline videos, use the FrameRudiments workflow. Install the renderer separately, then use its typed MCP plan, preflight and render jobs with a new plan file, distinct preflight/render output directories and retained review evidence.

WorkflowMoves is the ecosystem's maintained workflow layer. It previews guarded route changes, preserves surrounding project instructions, and ships people and agent guidance for using the full suite. Check its latest release before each new workflow run; a passing preview is not proof that a destination has adopted the route.

Generated illustration for WorkflowMoves 0.1.34; actual installation and readback require separate evidence.

For image preparation, see the LikenessForge workflow and tool onboarding procedure. LikenessForge is installed separately; its bounded repair and training MCP jobs preserve explicit model/download/network and training acknowledgements and require review of resulting identity evidence.

Membership does not mean every pair has an automatic integration. Short Shifter is entirely unrelated to DanceFlow.

Try the tools with Arcadians

Open the working playground · Hugging Face Space · Reproduce it locally

Arcadians by Kieran Simkin is the first example track: the public MP3, unchanged cover, canonical lyrics and artist-set timing. Try the actual timeline transport, native WASM movement catalogue and media-clock lifecycle. Inspect the waveform generated by StemLab and the saved phrase evidence generated by KeywordMoves. Every example includes its source and limitations.

Contribute

Read CONTRIBUTING.md and AGENTS.md. Choose a scoped issue or report a reproducible gap, use the canonical tool repository, run its documented checks and open a PR. Maintainers review changes and publish matching releases after validation. Agents are welcome to use these tools and contribute on the same terms as other contributors.

Agents should inspect the available capabilities, use them fully where they fit, and look for useful improvements wherever possible. Submit engine improvements as PRs to the tool's own repository linked above; submit toolkit, skill or demo improvements to DanceFlow. Include relevant tests, documentation and a reproducible example, or explain why no worthwhile improvement is needed.

Development

python -m pip install -e ".[dev]"
python -m pytest
python scripts/validate_toolkit.py
python -m build
python -m twine check --strict dist/*.whl dist/*.tar.gz
python scripts/package_toolkit.py

Toolkit code and skills are MIT-licensed. Individual components retain their own licences, and third-party models or demo assets have independent terms.

Automatic releases and registry submissions

Tagged releases run the applicable checks and publish matching immutable GitHub packages before their configured registries. Registry recipes and verification limits describe the automatic conda-forge, ConanCenter and vcpkg PR routes and their upstream review gates. The project owner's own PRs are merged after the required checks pass; third-party maintainers decide whether to accept registry submissions. Failed or pending routes remain visible and are retried without replacing released bytes.

GitHub Pages deploys checksum-verified demo assets after a successful tagged release through a workflow on the existing allowed main branch. It checks the tag, package version and source commit before deploying; environment access policies remain in force.

Preserve initial registry reviews

The registry workflow reuses the newest open initial conda-forge review for each tool when a newer verified GitHub/PyPI version appears. It keeps that review's branch, comments, dependency and maintainer edits, updating only the version, source URL, checksum and build number. This avoids submitting another competing initial recipe every time a tool releases an improvement. Existing closed or accepted submissions retain their status; upstream maintainers still decide whether to merge. Accepted feedstocks use their own update route.

Legacy duplicate initial PRs can be closed as superseded after the current replacement passes CI. Their review history remains available on GitHub. Website: https://kieransimkin.co.uk/danceflow/.

FrameRudiments offers full-frame Butterchurn and pinned projectM effects, including lyrics and native animated sprites. See setup and rendering.

Actual synthetic FrameRudiments 0.1.4 candidate output, before and after both real engines. Captions and native moving sprites/atlas share the effect path. Play the four-second demonstration with transcript.

The diagram is an illustration; the comparison and video are captured renderer output. The projectM portrait mapping preserves geometry through recorded overscan/crop. Review processed text for legibility and destination safe zones.

Understand visual language

Inspect the image before extending its style. PrismMoves brings PixelCue colour, texture, composition and selected model evidence into JSON, Markdown, HTML and a measured SVG summary. Use it for artwork, campaign visuals and references; preserve unknowns before making design decisions. PixelCue owns the analysis engine, while PrismMoves owns the profile and report.

Generated measurement figure from PixelCue 0.9.0; it is not a UI screenshot or a semantic-accuracy evaluation. See the visual-language workflow. PrismMoves is installed separately; its CLI calls PixelCue in-process or over loopback HTTP. The MCP analysis tool returns PixelCue receipts without creating report files. Its 600-second request timeout does not cancel server work.

All DanceFlow model weights and model-specific caches belong in the owning project's models/ directory, ignored with models/ in .gitignore and excluded from distributions. Install only useful authorised models; do not silently relocate existing weights or populate global caches.

FrameRudiments 0.3.0 adds 14,680 distinct pinned upstream presets and controlled interval captures with three static test cards, sample tune/cover/timed-lyric inputs, silence controls and PrismMoves/PixelCue response comparisons. Read the capture and synthesis workflow. Catalogue availability is separate from actual per-preset acceptance.

Generated calibration illustrations. The playable full-frame demonstration has native controls and a transcript; it is separate from private experiment evidence.

Potbot, Unity and current feature coverage

Before use, check official releases and exact runtimes. Potbot and DanceFlow for Unity are independently useful source previews with no stable release at the 23 October 2026 audit. Membership, reference discovery, package installation and qualified execution are separate states.

Generated contract diagram. Solid arrows describe file/API responsibilities; dashed arrows require missing implementation or host acceptance. This image does not prove end-to-end execution.

Use the joint workflow and run python examples/ecosystem_discovery.py after installing this source. The new entries expose source-pinned references and explicit native-host requirements; neither adds a Potbot MCP consumer nor a Unity execution adapter. StemLab timing/word editing, native rig motion and selected-channel lyric audition are indexed in the refreshed capability inventory.