🌼 DaisyChain

Script and workflow manager for DaVinci Resolve

Goals

Make DaVinci Resolve/Fusion scripting fast, fun, and reliable
Write scripts with full autocompletion, and in-IDE documentation
Turn your one-off script automations into ongoing organized workflows
Experiment with Resolve scripts safely, on an unbounded undo tree
Share your workflows and use those that others have shared

Status

Experimental ⚠️ (Use caution, most script executions are irreversible!)

Language	Resolve API	Fusion API	Tests
Python	✅	🚧	❌
Lua	❌	❌	❌
JS/TS	❌	❌	❌
Rust	❌	❌	❌

Why?

Blackmagic's API for Resolve and Fusion is extensive and powerful, but woefully undocumented and behind-the-times architecturally. From locking users into antiquated Python versions, to non-existent (or silent) errors/exceptions/tracebacks, to documentation in a .txt file without any IDE / language-server support, scripting is not fun.

DaisyChain makes it so you can use the Resolve/Fusion API from any language you like (Python 3.6-3.12+, Lua, JS, TS, Rust), in any execution environment you like (terminal, localhost web app, ...) while retaining full in-editor documentation stubs and types.

It does this by hosting a remote procedure call server within Resolve, and defining a serialized interface for the entire API surface. (The Resolve API itself is actually an RPC server, so this is really like an RPC of an RPC. While a little bit ridiculous, this allows us to write and use the API in a familiar way with full documentation)

Types, documentation, and RPC connection is written in each client language, see the status of your favorite language in the Status table above.

Script use-cases

🚧 A Chromium / WebKit extension to "Download to Media Pool" from internet, called CreditsDue
🚧 A script to copy from a current-timeline-item metadata field to clipboard on a timer, called CopyCat
🚧 A more accurate transcription workflow than Resolve's trashy one, using OpenAI Whisper
🚧 Removing silence from a clip automatically (via ffmpeg or some loopback)
🚧 py-auto-gui madness to automate repetitive edit tasks (macros, basically)

Audience

Jonny mostly, for now
Later, lazy video editors with coding experience
Later, production houses with workflow programmers on staff

Installation

Set Resolve External Scripting to Local or Network
Install Python
Ensure that the PYTHONHOME env variable is set at system level, to point to the directory where you will find the python3.exe, or symlink that path with /usr/local/bin/python3
Copy the DaisyChain Host script into Resolve's watch folder for scripts macOS
```
cp "./python/DaisyChain Host.py" "/Library/Application Support/Blackmagic Design/DaVinci Resolve/Fusion/Scripts/Utility"
```
Windows 📁 "%AppData%\Blackmagic Design\DaVinci Resolve\Support\Fusion\Scripts\Utility"

Linux 📁 "~/.local/share/DaVinciResolve/Fusion/Scripts/Utility"
Open
Close and open Resolve if it was open

Roadmap

Milestone 1 Python server and client

Python RPC Server
Python RPC client with Resolve API coverage, types, and docstrings
Python RPC client with Fusion API coverage, types, and docstrings

Milestone 2 multi-language

JS RPC client with API coverage, types, and documentation
Lua RPC client with API coverage

Milestone 3 reversibility & safety

RPC Server should
- Validate all input types before executing any function in the API
- Retain an infinite undo-tree (with timestamps) for reverting to previous states

Milestone 3 user interface & sharing

Build a workflow scheduler & package/script manager UI and package manager
Determine if you even need a workflow scheduler UI or if Warp/tmux/nu is just fine?

Contribution

Feel free to report bugs / issues, but do not expect any serious support on the Issues page.

If you would like to contribute a client package for your favorite language, or a suite of tests for existing code, please feel free to submit a pull request!

Technical notes

Since it seems like the Resolve API is designed for syncronous execution, let's stick to having syncronous execution of the RPC server, so that we can't overload the integration API. With that said, it would be worth doing some testing to see what happens with multiple servers running at the same time, and what the maximum refresh rate / transaction rate of the API is.

jonnyhyman/daisy_chain