MyPart Opens AI Music Matching Tools to Developers

MyPart has expanded developer access to its AI music matching and search tools through Replicate and Model Context Protocol, aiming to help apps and agents pair existing songs with visuals, prompts and user context. The rollout builds on Songhunt, an adaptive gaming experience launched with Epidemic Sound and Overwolf.

Why it matters: - MyPart is pushing music search and matching into AI workflows, which could make it easier for developers to add soundtrack selection, recommendations and music discovery to apps. - The tools focus on existing human-made music, not AI-generated tracks, which matters for creators, platforms and rights-aware commercial use. - The release extends MyPart’s music intelligence into gaming, video, streaming and personalized listening experiences.

What happened: - MyPart announced broader developer access to its music intelligence through Replicate and AI Model Context Protocol, or MCP. - The company launched two music intelligence models on Replicate: Visual-to-Music Matching and Prompt Song Search. - MyPart Labs also released general MCP AI music agents for conversational and automated creative workflows. - The rollout builds on Songhunt, an in-game adaptive music application on Overwolf developed with Epidemic Sound.

The details: - Visual-to-Music Matching suggests contextually relevant music for an image, video or collection of visual files, with an optional text hint to refine results. - Prompt Song Search finds songs from descriptive prompts covering mood, genre, instrumentation, lyrical themes, writing style and other creative criteria. - Both models return ranked recommendations from MyPart’s partner catalogs. - MyPart says the catalogs include real, human-made music rather than AI-generated tracks. - The MCP agents are designed to interpret music, lyrics, visual rhythm, scene transitions, audience context and user preferences. - Use cases include short-form video scoring, soundtrack selection for ads, presentations and websites, and personal music assistants across streaming platforms. - MyPart’s technology analyzes musical characteristics, vibe, key moments, structure and lyrics, then connects them to visual and textual context.

Between the lines: - The company is positioning music matching as infrastructure, not just a consumer feature. - Replicate and MCP give MyPart a distribution path into developer tools and agent ecosystems, where demand for contextual media intelligence is growing. - Songhunt serves as a proof point for real-time use, showing the system can react to changing on-screen action instead of relying on static playlists. - Epidemic Sound’s involvement signals that licensed catalogs remain central to adaptive music experiences, especially for creators and streamers who need shareable content.

What's next: - MyPart says developers, platforms and creators can use the new access points to integrate image and video soundtracking, contextual recommendations and music discovery into products. - Songhunt is meant to demonstrate how the matching technology can work in other real-time environments. - The company is directing users to access information for Replicate and MCP. - Additional information on Songhunt is available through the Overwolf app download and Epidemic Sound’s partnership announcement.

The bottom line: - MyPart is turning its music-matching technology into a developer-facing layer for AI agents, with gaming as the clearest example of what that could enable.