Video Editing with GPT-6 Astra in DaVinci Resolve via MCP

X postAlexDemo

Video editor Alex shares his experience using GPT-6 Astra connected directly to DaVinci Resolve via the Model Context Protocol (MCP). The model handles repetitive tasks such as syncing, rough cuts, and punch-in zooms, while the editor acts as a creative director overseeing corrections.

Key points

  1. Direct integration of GPT-6 Astra into DaVinci Resolve via an MCP server minimises reliance on full GUI computer use.
  2. Structured skill-based workflow: initial syncing of multi-camera footage, first automated cuts, and a dedicated cleanup pass.
  3. Automated standard editing patterns include punch-ins/zooms on talking heads and two-frame audio crossfades at cut points.
  4. Dialogue editing remains imperfect: Astra occasionally trims word endings or leaves pauses, requiring dictated audio fixes.
  5. Resource consumption: Editing an hour-long VOD down to 20–30 minutes uses around 10% of maximum plan usage on Astra Light.

Why it matters

Connecting LLMs directly to professional desktop NLEs via MCP demonstrates how AI agents can reliably perform background timeline manipulation instead of relying purely on screen capture and mouse simulation.

Analysis

Alex provides a grounded account of AI-assisted editing, steering clear of 'one-click edit' hyperbole. He notes that total editing time remains similar because supervision, playback reviews, and dictated fixes are still necessary, but tedious mechanical work is eliminated. While the DaVinci MCP provides fast execution, it currently lacks access to certain menus and audio effects, prompting the model to fall back on computer use. The approach highlights how agentic protocols enable background multitasking for content creators.

Original post

Alex

@The_Alex · 10 September 2026

GPT-6 Astra editing in DaVinci t.co/zVPIT9KWvx t.co/SkwSq3q78M

1k likes37 replies1.3k bookmarks268.3k views

View on X
More posts in the thread (10)
  1. @carlosrodera_ Editing down an hour VOD into a 20m video takes about 10% of my max plan usage. Although I am still learning and the better I get at promoting and creating skills the cheaper it will get
  2. @carlosrodera_ Also yeah so far just Astra Light
  3. This is a 1 hour VOD edited down to ~30 minutes
  4. @wohe_99 Soon yes. They are so close to being good enough to share. What I will say tho is to just use your voice and tell Astra what you want and then turn that into a skill for your own content. Then refine from there
  5. @GodsBoy7777 Trying this today! Been hearing this about Astra and that Light can be pretty hungry
  6. @KatotterFGCU Does Lightroom have an MCP? DaVinci does so it allows Astra to plug directly into it
  7. @DheerajShah Davinci has an MCP so Astra plugs directly in. Wicked fast and less computer use
  8. @jes_astro Video will be out today!
  9. @sadat_codes Will be live today!
  10. @NealWiser Yes but still manually watching through and it does a great job so far
Selected replies (5)
  • @AdranMontenegro @The_Alex Cuando pueda elegir por mi, las tomas que sirvan en un documental y descartar las que probablemente no usaré, ahí si creeré que es una gran herramienta de ayuda.
  • @carlosrodera_ @The_Alex Amazing, how many tokens / % of weekly subscription it takes per let's say 10/20 minutes? Do you use just one model, or do you delegate different tasks to different models? What I do is, for the plan, I use Astra, and for the work of delivery, I use Gemini 3.8 Flash. And
  • @NealWiser @The_Alex Are you using it to choose selects? Building a rough cut is incredibly useful, but if the selected shots are bad, what’s the process then?
  • @wohe_99 @The_Alex can you share the skills you're using
  • @carlosrodera_ @The_Alex Sounds great. Around 10% of weekly usage for a 1-hour raw video, without needing any complex workarounds, sounds perfect. Thanks for sharing!

Mentioned links

Summary generated by AI (Gemini 3.8 Flash, 27 September 2026). It may contain errors – the original source is authoritative.

Related entries

Similar in content, found by the AI search.

  • X post:el.cine

    Automating Video Workflows with GPT-6 Astra and Filmera MCP

    Creator el.cine demonstrates how connecting GPT-6 Astra to a Filmera MCP server automates video production workflows. Through natural language input, the model generates canvas nodes, prompts, character sheets, and scene settings.

    1.2kbookmarks100.3kviews

    AI & LLMs· Demo

  • X post:Matthew Berman

    Vibe Editing: Prompt-Driven Video Editing with DaVinci Resolve and OpenAI Codex

    Matthew Berman highlights a guide by Brian (@ForwardEditor) outlining how to start 'vibe editing'. The workflow integrates DaVinci Resolve with OpenAI Codex via a Model Context Protocol (MCP) connector to direct video editing iteratively using prompts and timestamped corrections.

    314bookmarks38.6kviews

    AI & LLMs· Tutorial

  • Video

    Video:CURT

    Recreating Product Launch Videos with GPT-6 Astra, DaVinci Resolve, and Blender

    CURT demonstrates an end-to-end workflow using GPT-6 Astra to clone and customise product videos. Using Codex, Hyperframes, and DaVinci Resolve, reference footage is rebuilt with frame-level control. The process integrates HeyClicky for voice-driven prompt engineering, MagicPath for visual canvas annotations, Diffusion Studio for UI motion, and Blender for 3D asset generation.

    AI & LLMs· Tutorial

  • Video

    Video:CURT

    GPT-6 Astra Computer Use: Five Practical Workflow Demos

    Curt showcases five real-world use cases for GPT-6 Astra's computer use feature within Codex, demonstrating how to automate tasks and test tools across desktop applications that lack APIs or connectors.

    AI & LLMs· Tutorial

Back to the knowledge base

Let's talk about your project

Get in touch

Locations

  • Mattersburg
    Johann Nepomuk Bergerstraße 7/2/14
    7210 Mattersburg, Austria
  • Vienna
    Ungargasse 64-66/3/404
    1030 Wien, Austria

Parts of this content were created with the assistance of AI.