# Slideshot

> AI agent that records polished product demo videos by driving your web app from a natural-language prompt and returning a ready-to-use MP4.

Slideshot is an AI-driven product demo video tool currently in alpha. It takes a natural-language description of a product flow, drives a browser through the target web app, and returns a polished MP4 — complete with optional intro cards, background framing, blur effects, and keyboard shortcut overlays — without any manual recording or editing.

## What It Is

Slideshot sits in the category of automated demo generation tools, distinct from screen recorders like Loom or ScreenStudio. Instead of a human manually capturing a walkthrough, Slideshot's agent executes the described flow in a browser — including authenticated product areas — and produces a launch-ready video asset. The tool is designed for product teams who need repeatable, up-to-date demo videos as their UI evolves.

## How the Agent Workflow Works

The core loop is three steps: describe the flow (product URL + goal written like a PMM brief), let Slideshot drive the app in a browser, then download the finished MP4. The agent handles navigation, interaction, and post-processing automatically. Runs are asynchronous — most complete in under 5–10 minutes, though longer demos can take 20+ minutes.

Slideshot supports saved credentials so the agent can sign into target apps and record real in-product workflows, not just public marketing pages.

## Integration Paths

The docs describe three ways to connect:

- **MCP server** — Add Slideshot to Claude Code, OpenAI Codex, ChatGPT, Cursor, or any MCP-capable agent. The agent discovers Slideshot tools and drives recording from natural language.
- **CLI** — Install the `slideshot` npm package for scripts, CI jobs, and scheduled automation via `npx slideshot runs create`.
- **HTTP API** — Call `POST /v1/agent/runs` directly with any HTTP client using an API key; poll for status and download artifacts when the run succeeds.

## Video Customization Options

Each run supports a range of visual controls:

- Intro/title card templates for branded demos
- Solid or gradient background canvas framing the recorded UI
- Configurable output dimensions and inner content scale
- Email blur to hide private identifiers during recording
- Keyboard shortcut overlays surfaced as on-screen annotations

## Why It Matters for Agent-Driven Teams

The homepage explicitly frames the API as a core product promise: "launch demos should be generated from workflows, agents, and automation, not only from a human opening a recorder." This positions Slideshot as infrastructure for teams that want demo video generation embedded in CI pipelines, changelog automation, or agent-driven launch workflows — not just a faster manual recorder.

## Current Status

Slideshot is currently in alpha, as indicated by the `[alpha]` label on the homepage. It is live on Product Hunt and accessible via `app.slideshot.ai`. The API, CLI, and MCP server integration paths are all documented and available.

## Features
- AI agent drives browser to record product flows
- Natural-language goal input
- Authenticated web app recording with saved credentials
- Polished MP4 output with no manual editing
- Optional GIF export
- Branded intro/title card templates
- Solid or gradient background canvas framing
- Configurable video and content dimensions
- Email blur for privacy during recording
- Keyboard shortcut overlays
- Asynchronous run API with polling
- MCP server integration for AI agents
- CLI via npx for scripts and CI
- HTTP API for direct integration
- Brand asset extraction endpoint

## Integrations
Claude Code, OpenAI Codex, ChatGPT, Cursor, MCP-compatible agents

## Platforms
WEB, API, CLI

## Pricing
Paid

## Version
alpha

## Links
- Website: https://slideshot.ai
- Documentation: https://slideshot.ai/docs
- EveryDev.ai: https://www.everydev.ai/tools/slideshot