Syllabus to Series
A syllabus becomes a run of short modules, storyboarded week by week for faculty to sign off.
The research studio for video at scale
Storyboards, professional voice cloning and generated media, with edits down to the second. For universities, training teams and products.
Universities
A full course, without a film crew.
Hand us a syllabus and lecture notes. We storyboard each module, narrate it in your lecturer’s cloned voice, and next term we update only what changed.
Get startedThe Lecturer’s Voice
Narration in your lecturer’s own cloned voice, with their permission, and no more hours in a booth.
Every Language on Campus
One master module, re-voiced for the languages your students learn in, with on-screen text re-timed to fit.
Training teams
Training that keeps up with the job.
Onboarding, compliance and product training, made from the documents you already have. When a process changes, we change that step and leave the rest.
Get startedSOP to Walkthrough
A written procedure becomes a step-by-step video, with the screen doing the showing.
Knowledge Checks
Questions placed at the moments that matter, each tied to the section it tests.
Update One Step
Policy changed? Re-voice and regenerate that step. Every other approved minute stays as it is.
Products and channels
Your product and your channel, on film.
The same pipeline, pointed at SaaS launches and faceless channels, and made as often as you ship.
Get startedLaunch Films
A cinematic cut for the day it ships, built from your actual interface.
Ad Sets in Every Ratio
Vertical, square and wide, recomposed for the frame instead of cropped to it.
Faceless Series
One locked house style across every episode, so episode 90 still looks like episode 1.
Three models under every frame.
Indexes every frame, sound and caption by what it means.
Our in-house understanding models reason over a whole video, not one frame at a time.
Agents act on that reasoning and change only the word, shot or second that failed.
Research
The problems we work on.
Generating a video is the easy half. Knowing whether it is any good, and fixing it without starting over, is where our research goes.
QA agents in a loop
Reviewers that each watch for one failure, such as continuity, pacing or legibility, and keep sending work back until it passes.
Precise edits
Changing one word or five seconds so it still matches the frames on either side, leaving the rest of the video untouched.
Reasoning over long video
Models that hold a story in mind across minutes, and can point to the moment behind each conclusion.
Measuring taste
Turning “this looks good” into a signal an agent can act on, so aesthetic quality holds at volume.
Work with us
Bring us a brief.
Send a syllabus, a training manual or a product brief. We’ll tell you what we’d make from it.
hello@bastlerguild.com
FAQ
Common questions
What is Bastlerguild, exactly?
A media research company and a studio in one. The research builds the models and agents; the studio uses them to make video.
Who is this for?
Universities and training teams first, then product teams and faceless channels. Anywhere a lot of video has to stay accurate over time.
How does storyboarding work?
Before anything renders, you get every scene, shot and line of narration on one board. Changes are made there, and generation starts once you sign off.
Can you clone a lecturer’s or trainer’s voice?
Yes, with that person’s permission. A recorded sample becomes a voice that can narrate a whole course and be updated later without a new session.
What do you mean by a precise edit?
Change one word and only that word is re-voiced. Replace one shot and only that shot is regenerated, matched to what is around it. QA agents then review the change before it goes back in.
Are the video understanding models your own?
Yes. We build our video understanding models in-house, including the ones that reason across a whole video.