A Finnish video AI company founded in 2015 out of the University of Oulu’s computer vision lab, offering a multimodal video understanding engine that analyses speech, visuals, audio, on-screen text, faces, emotions, and structure in a single pass. Two products sit on the same engine: a REST API (JSON-RPC 2.0) that returns structured JSON metadata across 38+ detection types—speech-to-text, speaker separation, face/object/brand detection, IAB content classification, sensitive-content flagging, scene descriptions, mood and emotion—with support for 50+ formats up to five hours or 7GB and EU data processing; and Valossa Assistant, a conversational front end for non-developers that accepts an uploaded video and returns timestamped answers, clips, transcripts, chapters, and content-safety reports, with MCP support so Claude and Cursor can call it directly. Assistant pricing starts at €14.90/month with a 7-day free trial (30 minutes of analysis plus 100 credits), and €29 credit packs; API high-volume usage, Ad Scout, Autopreview, and private deployment require contacting sales. Public customers include MTV Finland and Cineverse. The company has raised roughly $2.65M total, with no new round since 2017, and has about 17 employees.
**Pros & Cons**:
Pros:
Multimodal single-pass processing across speech, visuals, audio, on-screen text, and faces is more efficient than chaining separate transcription, vision, and classification services; 38+ detection types including IAB content classification and sensitive-content flagging cover adtech and brand-safety use cases that generic video APIs often skip; EU data residency suits European broadcasters and regulated buyers; real public customers (MTV Finland, Cineverse) give it more production validation than many newer video AI startups; Assistant provides a non-developer front end on the same engine, with MCP support for Claude and Cursor; supports 50+ formats up to five hours or 7GB per file.
Cons: **Small company with thin funding**—roughly $2.65M raised, no new round since 2017, about 17 employees, so long-term independent growth is uncertain compared with well-funded competitors; independent third-party user reviews are absent—credibility rests on public customer case studies rather than G2/Capterra/Trustpilot ratings; API pricing is opaque—Assistant is priced from €14.90/month, but high-volume API usage, Ad Scout, Autopreview, and private deployment require contacting sales, making cost forecasting difficult; MCP support was added in 2026 rather than being native, so it should not be described as MCP-native; name collision with Valsoia, an unrelated Italian food brand with poor Trustpilot ratings, creates search confusion.
Who is it for?:Broadcasters and media companies that need automated metadata, transcription, and content classification for large video archives; adtech and brand-safety teams requiring IAB classification and sensitive-content flagging; European organisations that need EU data residency for video processing; non-technical users who want to query video content conversationally through the Assistant.