The AI video editor built for long-form.
Ten raw minutes in. Your timeline out.

The model writes. NeoMotion does. Searching stock, generating and cutting out images, colour-matching, casting the right preset to the right sentence, keyframing a timeline, and exporting are app operations — not text generation. Swap the model out and that pipeline still works. That's the part we built.

Yes. NeoMotion runs a local MCP server on your machine, so Claude Code, Claude Desktop, Cursor, or anything that speaks MCP can operate the app directly — create a project, import media, compose scenes, author a custom motion graphic, render. You pay your provider; we take no cut.

Not to start. Voiceover and English transcription run locally on your machine — no key, no bill. For script-writing, image generation, and non-English transcription you either paste your own keys (OpenAI, Gemini, Replicate) or connect the AI subscription you already pay for over MCP. Either way, we never mark up inference.

Yes — that's half the point. Start from a one-line idea, or paste a finished script and the crew keeps it word for word. It researches, writes, voices, gathers visuals, and builds the composition without you filming anything. And because rendering happens on your machine, there's no credit meter — the economics that break most faceless workflows don't apply.

It can't. Every caption and cue is anchored to your transcribed words — down to the word. The AI proposes; you approve; nothing renders without you.

No. NeoMotion runs on your machine (Apple Silicon Macs on macOS 12+, or Windows 10/11). Editing, rendering, and English transcription happen locally — multi-gigabyte footage never gets uploaded, because there's nowhere for us to upload it to. AI features send only the little they need — a search phrase, a thumbnail — to the providers you choose, never to us.

Any laptop from the last few years. The floor is 4 CPU cores, 8 GB of RAM and about 15 GB of free SSD space. The sweet spot is 8 cores and 16 GB — that's where exports feel quick and you never have to touch a setting. On Mac you need Apple Silicon (M1 or newer) on macOS 12 or later; on Windows, 10 or 11, 64-bit. There's no Intel Mac or Linux build.

No — and that's on purpose. NeoMotion renders frames on the CPU so the same project produces identical frames on every machine, which is exactly what lets us verify a scene before you ever see it. No CUDA, no minimum VRAM, no graphics card shopping. Your GPU still helps where it's free: hardware video encoding on the final file, and 3D, map and shader layers while you're editing. If you want faster exports, more CPU cores is the lever — nothing else comes close.

Around 800 MB to download and about 2 GB installed, because NeoMotion ships its own renderer, ffmpeg and browser rather than making you install them. The optional local voices and transcription models add up to 2.9 GB more, and only if you switch them on. Your footage costs nothing extra: video and audio are edited where they already sit, never copied into a library. Plan for 15 GB free, or 50 GB if you cut long recordings.

It's your computer. Your CPU, your electricity, our software. There's no server in the loop, so there's nothing for us to bill and no meter we could run even if we wanted to. The real trade is time: a long 4K export ties up your machine instead of a render farm. Worth it, around version nine.

Founding members keep beta pricing for life. Everyone who joins after launch pays the standard rate. That's the trade: you help us polish it, you keep the founding price.