Now Announcing

Performm Vision

Find the moment
you're in search of.

AI Search that makes the contents of every video, audio and image file searchable.
Transcripts, faces, objects and scenes, across 90+ storage providers.

Video·Audio·ImageOne intelligence layer
90+Storage providers
Universal SearchFind and create
The problem

The knowledge inside your media is invisible.

Transcripts, on-screen text, faces, objects and scenes stay locked in the file. Cloud-only AI tools are a non-starter for private, legal or industrial content. Teams store petabytes they cannot actually look inside.

What it does

Understand. Search. Explore.

One pipeline turns raw media into something you, and your agents, can actually query.

💬
Understand
Speech-to-text with timestamps, in-frame OCR, object, face and logo detection, scene and shot segmentation, plus summaries and keywords for every file.
🔍
Search
Full-text and semantic vector search across every connected store. Ask in natural language and find the moment by meaning, not by filename.
▶
Explore
An interactive timeline with visual overlays takes you to the exact frame or phrase. Local-first and private by default, on your infrastructure.
Capabilities

Everything Vision sees.

A full media-understanding stack, running across all 90+ providers Performm already connects.

🎤
Speech & audio
Timestamped transcription, speaker cues and audio keywords.
📄
Text in frame
OCR reads slides, signage, captions and documents on screen.
👤
Objects, faces & logos
Detect and tag what appears, frame by frame.
🎬
Scenes & shots
Automatic segmentation into searchable scenes and shots.
🧮
Semantic search
Vector search by concept and intent, across every store.
📝
Summaries & metadata
Auto summaries, keywords and structured metadata per asset.
Agent-native

Built for AI agents, over MCP.

Vision exposes every result through the Performm MCP server. Claude, Cursor, Agentforce and any MCP client can search and reason about your media content directly. Storage becomes something agents can read, not just hold.

◉ Claude ◉ Cursor ◉ Agentforce ◉ Any MCP client
# agent calls Performm Vision over MCP
vision.search(
  query="CEO on stage, product reveal",
  modalities=["video","audio"],
  scope="all-connected-stores"
)
→ 7 moments across 3 providers
→ deep-linked to the exact frame
Two ways to run it

A feature in Performm.
A product you can license.

Inside Performm
Native to the platform
Vision ships as a capability of Performm. Switch it on across the storage you already connect, with nothing new to integrate.
  • Works across all 90+ connected providers
  • Exposed through the Performm MCP server
  • Supercharges Performm Search out of the box
Standalone license
The Vision engine, on its own
License Vision as a standalone product, powered by Amove, and embed it in your own product, archive or pipeline.
  • Local-first, private deployment
  • API and MCP access to every result
  • Full source rights available for enterprise
Who it is for

Built for libraries that never stop growing.

🎬
Media & post-production
Find any shot, line or face across petabytes of footage.
🏫
Training & knowledge
Make internal video and audio libraries instantly searchable.
🛡
Security, legal & compliance
Search private archives locally, with nothing leaving your walls.
🤖
Industrial & robotics
Index sensor, camera and inspection footage at scale.

See inside everything you store.

Put Performm Vision on your media and let your agents do the looking.

Request a demo