Montage
Setup Guide
This guide is for invited desktop and terminal users. Developers can use the repository docs, but most users should start in the desktop app: open Settings, sign in, add the provider keys they need, then run a small project before using paid generation.
Step 1
Sign in
Use the in-app OpenAI or ChatGPT sign-in before asking the agent to work.
Step 2
Add keys
Open Settings → Advanced → Provider Keys and add only the providers you plan to use.
Step 3
Test locally
Import a short clip, run indexing, ask for a simple edit, then render/export.
Provider Keys
OpenAI or ChatGPT
The editing agent and chat-driven project work.
Sign in from Montage when prompted, or use an OpenAI API key if that is how your access is configured.
Get keyOpenRouter
Generated video or generated B-roll requests.
Create an OpenRouter account, add credits, create an API key, then add it in Montage Settings.
Get keyDeepgram
Cloud speech-to-text transcription.
Create a Deepgram project and copy an API key into Montage Settings.
Get keyHugging Face
Gated model downloads used by some diarization workflows.
Create an access token on Hugging Face and accept any model license required by the workflow.
Get keyAnthropic
Optional premium topic labeling and analysis paths.
Create an Anthropic Console key and add it only if your workflow needs it.
Get keyX
Trend/context reads where enabled.
Create a developer bearer token and add it only if you use X trend tools.
Get keyPython and indexing
Montage uses a bundled Python workspace and uv for media indexers. If indexing fails, open the app logs and send the error text with the project type, media duration, and which indexer was running.
Logs and support
From the desktop app, use Help or Settings to open the log folder. Include recent logs when reporting a crash, failed provider call, failed indexer, or export problem.
Contact tmbuwa09@gmail.com for access and support.