Hand over raw footage and receive a first cut closer to your own editing taste: that is Selendis’s proposition for Archon. The company describes the model as the foundation for personal AI video editors. Access remains at the early-access and waitlist stage.
What learning an editor’s taste is meant to do
Selendis says Archon adapts to a creator’s preferences and can edit video content in minutes, delivered through an AI-native editor for shared projects. The timing is a company claim, rather than a measured turnaround for every kind of footage.
Editing taste involves more than choosing an attractive shot. How long to hold a glance, where to cut a speaker and how to pace the scenes can turn the same material into very different stories. I would therefore judge this kind of editor by how well subsequent choices reflect corrections, alongside the quality of its first rough cut.
The waitlist is not an access commitment
The starting point is joining Selendis’s waitlist and reserving a username. Its November 2, 2026 terms describe a free pre-launch program for people aged 18 and over. Points and rank can inform access decisions, but confer no entitlement. A general opening date and paid-service pricing are not specified.
For an urgent production, a waitlist registration cannot replace an available editing environment. Actual account access comes first; assessing how well the tool preserves a team’s choices using real footage comes after that.
Separate project processing from model-training choices
The privacy policy covers project media, scripts, edits and assistant conversations. Processing supports service operation, project syncing and adaptation to an individual or shared project’s direction. Collaborators can see names, usernames and activity in shared projects. Hosting and AI-processing providers are also involved, so the desktop app should not be assumed to process everything locally.
Contribution to model training has a separate setting. The policy says it starts off in the EU, EEA, UK and Switzerland, and when a country is unspecified; it starts on elsewhere. It can be changed in Settings. Switching it off stops new training use, while models already trained can retain what they learned.
A team’s first decision is what it can share
Deleting a project differs from deleting every associated record. The policy allows some data to remain with an account, and shared work can remain in collaborators’ projects. Teams are also responsible for content rights and any optional connected-account publishing permissions.
Archon’s interesting aim is to reduce how often creators must explain the same editing preferences. Its value for a team also depends on the data choices: which footage may be shared, who may edit it and whether training contribution is enabled. Making those decisions before uploading the first folder is part of retaining creative control.
