Launchscene

npx add-mcp 'https://launchscene.app/api/mcp'

MCP server · Free while in beta

Real launch references. Ready for your agent.

Your agent searches reviewed launch scenes, reads their timing, text, camera and sound, then builds something that belongs to your product.

Your agentconnecting…

0 MCP calls0 scenes read0.0 s

A real exchange with the library, replayed. Hover to pause.

19

tools

6

guided prompts

100%

reviewed scenes, always labelled

0

media files shared

Search like your agent does.

Describe a scene. Real results come back from the library, with the response your agent would read.

HookElorian3.9 s

HookElevenLabs4.2 s

HookSaaSWAVE1.5 s

HookSaaSWAVE1.4 s

HookSaaSWAVE4.3 s

HookSaaSWAVE4.1 s

Real results from the published library. The agent view is simplified; full responses add sources, verification and missing-data reasons.

From a brief to a film plan.

1

Find a direction

Search scenes by role, technique, duration and style. Compare contrasting references instead of copying one.

2

Read the measurements

Open a recipe: text timing, relative boxes, camera movement, audio and frames, each with its source and limits.

3

Make it your own

Plan a structure, adapt a recipe, compute a beat grid. Every extrapolation is labelled as one.

Guided prompts

/launchscene-film

Build a launch-film plan from measured, contrasting references.

/launchscene-hook

Explore five contrasting openings for a product and promise.

/launchscene-like

Adapt the structure and motion logic of a chosen reference.

/launchscene-reference

Turn a chosen reference into a frame-numbered brief for your build.

/launchscene-review

Discuss a storyboard in the context of corpus variation.

/launchscene-vertical

Adapt a wide composition to a vertical canvas.

19 tools, four jobs.

Start with three: search_scenes, search_videos and reference_brief. Results come back as images your agent can see. The rest go deeper: inspect, measure and keep.

search_scenes

Start here. Describe the scene in plain words; every result comes back as an image your agent can see, signed with the brand, the role and the time.

Example call

search_scenes({ query: "a cursor clicks Export and the file appears" })

Returns

• A signed image per result: key frame and progression
• Scene id, brand, role, start and duration
• Techniques and on-screen text
• image_url and a Launchscene link to share

Plug it in. 30 seconds.

Add the URL to Claude, Cursor, VS Code or ChatGPT and sign in with your Launchscene account. Free while in beta. See pricing

Settings → Connectors → Add custom connectorCopy

https://launchscene.app/api/mcp

Paste this URL as a custom connector in Claude (web, desktop or mobile), then sign in to Launchscene when asked.

Account-linked clients can save references to private boards shared with the website. See or disconnect them from your account.

Skills for your agent.

Free skills that teach Claude Code how to work with Launchscene: plan a shot from a real reference, then check your text timing before you render. They call the MCP, so you need a Launchscene account.

for s in launchscene launchscene-reference launchscene-timing-check; do
  mkdir -p ~/.claude/skills/$s && curl -fsSL https://launchscene.app/mcp/skills/$s.md -o ~/.claude/skills/$s/SKILL.md
done

Installs for every project in ~/.claude/skills. Run it again to update. Use.claude/skills instead to keep them in one repository.

Launchscene references

launchscene

Research, plan, adapt or review a launch film from measured references, with Remotion and HyperFrames exports.

.zip for claude.ai Read SKILL.md

Reference brief

launchscene-reference

Turn one launch scene into a frame-numbered REFERENCE.md and BRIEF.md before you build the shot.

.zip for claude.ai Read SKILL.md

Timing check

launchscene-timing-check

Check every text in your composition against real launches, render the frames that matter, propose scoped fixes.

.zip for claude.ai Read SKILL.md

What’s measured, honestly.

Every reference carries its source, how it was checked, and what is missing. Your agent knows the limits too.

Text boxes are observations

A box is the position of the text at a recorded OCR frame. Older analyses may only give a midpoint estimate.

Easing is a label, not a curve

Camera moves are measured; the easing class is an estimate of their shape, never the original keyframes.

Reviewed is not every pixel

A reviewed video was checked scene by scene, by a human or by an AI reviewer, and says which. Not every OCR box or beat was checked one by one.

Statistics describe, they don’t judge

Corpus numbers come with their sample size. They show how launches vary, not what is good.

Open to everyone

Give your agent real references.

Create a free account, add the server URL to your client and sign in when asked. No invitation needed while in beta.