More screenshots
Works with the AI Agents you already use
voxim
Turn a lesson plan into short, narrated, captioned animation clips for the parts that need a visual, plus a full video and a teacher's guide. Runs offline.
Free
voxim
Example session with this skill installed
The request
Here's my Grade 7 science lesson plan on the phases of the Moon (attached: teacher-notes.md). Make short animated clips for the parts that need a visual, with narration and captions.
- Read your context and instructions
- Compiled the voxim
I read your lesson plan and made 6 clips for the parts a visual helps most. The bell ringer, the lamp-and-ball activity (students should discover that themselves), closure and the exit ticket have no clips, on purpose.
Full video: film.mp4 (2:21), with captions in captions.srt
Clip Length Where it fits in the lesson
01-MirrorLight.mp4 0:16 After asking: does the Moon make its own light?
02-HalfLit.mp4 0:24 Straight after MirrorLight
03-TwoViews.mp4 0:37 After the "half is lit" point, before the activity
04-Shadow.mp4 0:26 Right after TwoViews: tackles the misconception that Earth's shadow causes phases
05-Sequence.mp4 0:21 End of direct instruction; replay at closure
06-Check.mp4 0:20 After the activity; pause on the question
TwoViews shows the Moon's orbit from far above and, side by side, the phase seen from Earth, with a day counter.
Added beyond your notes: "not to scale" labels, and the day counter (from your 29.5 days). Phases are shown as seen from the Northern Hemisphere.
guide.md has the clip order, chapter times for YouTube or your LMS, and these notes. Captions are burned into every clip.
Connects securely to your tools. The creator never sees your data.
About this skill
Voxim turns a lesson plan or teaching notes into short, narrated, captioned animation clips.
Give your agent your notes. Voxim finds the parts of the lesson where a visual genuinely helps (a process, a cycle, something spatial or invisible, a common misconception), animates them, and hands back:
- one short clip per concept, with captions burned in, ready for slides or your LMS
- the whole set joined as one film, plus a captions file (.srt)
- a teacher's guide saying where each clip fits in the lesson, with a check question
It also makes product explainers and narrated website walkthroughs. Want your own voice? Record the script on your phone and Voxim cuts your takes into the narration and times the animation to your speech.
Requirements
- An agent that can run shell commands: Claude Code, or Claude with code execution. Tested with Claude Code. Installers for macOS, Linux and Windows are included.
- About 1.5 GB of disk space.
- First run (one time, a few minutes): installs a private Python 3.13 environment in ~/.voxim/ and downloads the open-source Kokoro voice model (about 340 MB) from GitHub. You don't need to install Python or ffmpeg yourself.
- After that everything runs offline on your machine. No API keys, no accounts, and nothing about your lessons is sent anywhere.
Free and open source (MIT): github.com/timolumide1/Voxim
Requires ~1.5GB disk space. First run needs network for model downloads. Tested on Linux; macOS/Windows support via included scripts. Needs ffmpeg for burned captions.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed today
- Free with an Agensi account
Requires
Permissions
Allowed Hosts
File Scopes
Network access is only for the one-time install: uv (astral.sh), Python packages (pypi.org) and the open-source Kokoro voice model (GitHub releases, ~340 MB). Optional features download a speech recogniser for recorded narration (sherpa-onnx, GitHub) and a browser for website demos. After installing, Voxim works fully offline and sends nothing anywhere.