FrameThrower

Company profile:

Startup name: FrameThrower

Tagline: The world’s largest film-still library for inspiration, learning, and making movies faster.

Elevator Pitch: Explore the world’s largest library of film stills for inspiration, learning, and making movies faster. FrameThrower lets film and advertising creatives search thousands of films by lighting, lens character, shot size, colour and mood. Describe the shot you need, collect references in lookbooks, and export PDFs or image boards.

Built by senior filmmakers and advertising creatives, FrameThrower also gives builders eleven REST endpoints, an MCP server for Claude and ChatGPT, an npm SDK and a ComfyUI node with depth-map support. It brings real cinematographic references into creative tools and agent workflows. Free and paid plans are available.

Target Market: Filmmakers, directors, cinematographers, advertising creatives, film students, and developers building creative tools and AI workflows.

How will you make money?: Freemium access with paid Pro and Studio subscriptions, plus usage-based API credits for developers.

How much capital have you raised?: null

Website: https://framethrower.ai

City/Country:

AI-assisted summary:

FrameThrower is a cinematic-reference platform that lets creatives search a large catalog of film stills and organize them into visual development workflows, while also exposing the library through a REST API, JavaScript SDK, and Model Context Protocol (MCP) server. (Source: https://framethrower.ai/) Its relevance is that it treats film references not only as a browsing experience for directors and designers, but also as searchable inputs for production tools and AI-agent-assisted creative workflows. (Source: https://framethrower.ai/developers)

Problem and target users

Film, advertising, photography, and design teams often need to communicate lighting, composition, color, camera language, and mood before production, but conventional image search makes it difficult to search specifically by cinematic craft attributes. FrameThrower positions its product around searching frames by concepts, color, mood, composition, lens character, shot type, camera angle, setting, and visual style. (Source: https://framethrower.ai/developers)

The platform is aimed at filmmakers and other creative professionals building lookbooks, pitch materials, treatments, moodboards, and previsualization references; its paid Pro plan is explicitly positioned for working filmmakers and AI creators, while its Studio offering targets crews. (Source: https://framethrower.ai/pricing) For developers and creative-tool builders, the company also offers programmatic access to the same reference library through API endpoints and an MCP server. (Source: https://framethrower.ai/developers)

Product and solution

At the product layer, FrameThrower offers a searchable film-still library and moodboard workflow where users can find references and organize them by scene, sequence, or look. (Source: https://framethrower.ai/) The company also says users can upload a script to generate a scene-by-scene visual board intended for pitching or previsualization. (Source: https://framethrower.ai/)

• Natural-language search can return frames matching a described scene, mood, lighting setup, composition, or visual idea. (Source: https://framethrower.ai/developers)
• Image search and color search allow users to find visually related frames or frames dominated by selected color palettes. (Source: https://framethrower.ai/developers)
• Craft-based browsing can filter for attributes including shot type, lens character, time of day, camera angle, visual style, director, genre, and year range. (Source: https://framethrower.ai/developers)
• The MCP server provides four tools—search, craft filtering, similarity search, and frame-detail retrieval—for compatible AI clients, using OAuth 2.1 authentication. (Source: https://framethrower.ai/mcp)
• FrameThrower states that its MCP catalog spans 5,489 films and that its results include film, director, cinematographer, craft metadata, thumbnails, and deep links. (Source: https://framethrower.ai/mcp)

The company’s API design is notably reference-oriented rather than asset-oriented: its documentation says responses provide metadata, thumbnail URLs, and deep links rather than raw image bytes. (Source: https://framethrower.ai/developers) This distinction matters because FrameThrower’s terms state that it does not own or license the underlying film frames, and that users remain responsible for rights clearance for downstream uses. (Source: https://framethrower.ai/legal/tos)

Business model, pricing signals, and traction

FrameThrower combines freemium subscriptions with metered usage: its free tier includes limited daily searches, downloads, image searches, and moodboards, while Pro is listed at $14 per month or $7 per month when billed annually. (Source: https://framethrower.ai/pricing) Studio is sold as an annual per-seat offering for teams of three to ten seats, with a listed price of $115 per seat per year. (Source: https://framethrower.ai/pricing)

API and MCP access are metered by credits across plans; the pricing page states that a $20 top-up buys 20,000 credits, with standard read operations costing two credits and image-model operations costing more. (Source: https://framethrower.ai/pricing) The company also distributes an official JavaScript client, framethrower-ai, through npm. (Source: https://www.npmjs.com/search?q=keywords%3Aframes)

Expert take

FrameThrower’s strongest positioning is its attempt to make cinematography reference searchable by production-relevant attributes rather than treating image discovery as a generic moodboard problem. The combination of creator-facing tools and developer infrastructure could make it useful across individual filmmakers, studios, and software platforms.

The principal uncertainty is rights and workflow complexity: the product is explicitly a reference service, not a licensing layer, so professional users will need clear internal policies for how references are shared, transformed, or used in AI-assisted pipelines. The durability of the business will likely depend on catalog quality, search relevance, and whether its integrations become embedded in real production workflows.

Note: Information based on publicly available sources at the time of writing, and summarized by AI.