Last updated on March 2nd, 2026 at 12:25 pm
With Musiio closing its API service soon, many music platforms are facing a time-sensitive challenge: keeping their search and discovery workflows operational without disruption.
If your product, internal tools, or customer-facing experience rely on similarity search, replacing your search provider is more than a backend adjustment. Search directly impacts user trust, discovery quality, and product performance.
This guide outlines a practical way to migrate similarity workflows from Musiio to Cyanite; and how to use this transition as a product upgrade.
Don’t miss out the first part we did on this – focussing on the migration from Musiio’s to Cyanite’s Auto Tagging. Check it out below.
What changes when switching a music search provider?
Replacing a similarity search provider is not just a technical endpoint swap. Even if two systems both offer similarity search, ranking behavior, reference handling, and filtering capabilities can differ.
A smooth migration therefore focuses on:
- replacing the API endpoints
- validating search results internally
- ensuring the product experience remains consistent
Cyanite Search in one paragraph
Cyanite provides audio-based search via API for music libraries, streaming services, sync platforms, and music-tech companies.
Search workflows can be built using:
- Similarity Search (single-track reference)
- Advanced Search (add-on for multi-track similarity and advanced metadata filtering)
- Free Text Search (natural language search)
Similarity can be performed using:
- Your own track IDs
- MP3 uploads
- Spotify links
- YouTube links
- any of the above combined (Advanced Search only)
Step-by-step migration plan
Step 1: Start testing immediately (Spotify-based evaluation or test environment)
Before replacing your production similarity workflows, the first step is to test Cyanite’s search capabilities in isolation.
You can begin immediately by testing similarity search against Cyanite’s Spotify-based showcase database. This allows your team to:
- evaluate similarity quality
- compare ranking behavior
- test reference workflows (track IDs, Spotify links, etc.)
No full catalog setup is required for this initial evaluation.
If you want to test similarity search against your own full catalog, we can set up a dedicated test environment together.
To get started, create an API integration here:
https://api-docs.cyanite.ai/docs/create-integration/
Similarity Search documentation:
https://api-docs.cyanite.ai/docs/similarity-search
You can then:
- run similarity searches using track IDs
- test Spotify and YouTube links
- explore multi-track similarity
- combine similarity with filters via Advanced Search (per request)
If you would like to test Advanced Search (multi-track similarity, similarity scores, and metadata filtering), simply contact us at business@cyanite.ai and we’ll enable it for your evaluation.
Step 2: Identify your similarity search inputs
Most Musiio customers use similarity search in one of these ways:
- Searching similar tracks using a track ID from their own catalog
- Searching similar tracks using an external MP3 upload
- Searching similar tracks using a YouTube link
Cyanite supports all of these workflows and additionally supports Spotify links.
Your first step is to map your existing Musiio workflow to one of these Cyanite input types:
- Track ID (fastest and most stable)
- Audio upload (MP3)
- External links (Spotify or YouTube)
Step 3: Replace similarity workflows (real-time vs external references)
Track ID-based similarity (instant results – recommended for real-time use cases)
Using your own track IDs is the most stable and fastest approach.
This is ideal for:
- “Show similar” features
- user-facing discovery modules
- recommendation systems
- internal sync tools
With track IDs, similarity search operates in real time.
Cyanite supports up to 10 search requests per second, making it suitable for production-grade discovery experiences.
External reference workflows (results after analysis – MP3, Spotify, YouTube)
External references are useful for:
- searching your catalog using a client reference track
- brief matching
- creative mood board discovery
Cyanite supports similarity search using:
Track ID-based searches return results in real time. External references typically require a few seconds up to around a minute for analysis before results are returned.
Before switching production endpoints, we recommend validating ranking quality and relevance with a representative sample of your catalog.
Step 4: Upgrade with Advanced Search (instant results – multi-track similarity + filtering)
Once single-track similarity is stable, many teams extend their setup using Advanced Search, which acts as an add-on to Similarity Search.
Advanced Search extends similarity from a simple reference match to a controllable discovery layer:
- Multi-track similarity (up to 50 reference tracks)
- Similarity scores, quantifying how close results are in percentage terms
- Most Relevant Segments
- Custom Metadata Filters
- Up to 500 search results
Multi-track similarity is particularly powerful for:
- playlist generation
- “Discover Weekly” style workflows
- brief-based search where multiple references define a sound
Importantly, Advanced Search also allows you to combine similarity with your own metadata.
You can:
- search for tracks similar to a reference
- while filtering by internal tags
- or by metadata such as release date, territory, clearance status, new releases, or priority tracks (anything that you attach as a custom tag to your tracks)
This enables highly controlled discovery workflows that go beyond simple similarity replacement.
Step 5: Add Free Text Search (instant results – optional but high-impact upgrade)
While Musiio did not offer Free Text Search, Cyanite offers this feature, complementing Similarity Search.
Free Text Search allows users to search using natural language queries such as:
- “uplifting acoustic pop with female vocals”
- “dark cinematic tension build”
- “minimal piano with emotional atmosphere”
- “lofi beats for studying”
For music libraries and sync platforms, this can significantly improve:
- discovery speed
- usability for non-expert users
- onboarding experience
- catalog accessibility
Many teams migrate similarity first, then introduce Free Text Search as a second-phase upgrade.
Example migration timeline
Day 1:
Create an integration and test similarity with track IDs.
Day 2–3:
Replace similarity endpoints in staging and review results.
Week 1:
Go live with single-track similarity replacement.
Week 2+:
Add Advanced Search and optionally introduce Free Text Search as a product upgrade.
A note on migration
Although both Musiio and Cyanite offer similarity search via API, the underlying concepts and implementation details differ.
This means migration is not just a technical endpoint replacement. It requires a short evaluation phase to ensure alignment with your existing product logic and user experience.
In practice, most teams complete this evaluation within days, but it should not be skipped.
Final thought: replace or improve
Many teams use this moment to:
- strengthen their discovery experience
- introduce multi-track similarity
- enable Free Text Search
- modernize search workflows without building a large data science team
If your team is affected by Musiio’s shutdown, we’re happy to support you with migration guidance.
Get migration support
If you want support migrating from Musiio to Cyanite, you can:
- Fill out our Typeform and book a call
- Or contact us directly at business@cyanite.ai
Developer resources
- Create an integration:
https://api-docs.cyanite.ai/docs/create-integration - Similarity Search (Track ID / MP3):
https://api-docs.cyanite.ai/docs/similarity-search - YouTube link analysis:
https://api-docs.cyanite.ai/docs/analyze-youtube-tracks
FAQs
Q: Which similarity search API can replace Musiio?
A: Cyanite offers audio-based Similarity Search via API for track IDs, MP3 uploads, Spotify links, and YouTube links. Advanced Search and Free Text Search provide additional capabilities beyond Musiio’s feature set.
Q: Can I migrate similarity search from Musiio quickly?
A: Yes. Many teams begin by replacing track-ID-based similarity workflows first, as this allows real-time continuity with minimal product disruption.
Q: Does Cyanite support multi-track similarity search?
A: Yes. Multi-track similarity (up to 50 reference tracks) is available via Advanced Search. This is especially useful for playlist generation, brief-based search, and recommendation workflows.
Q: How can I test Advanced Search?
A: Advanced Search can be enabled for evaluation upon request. Simply contact business@cyanite.ai and we’ll activate it for your integration, typically within one business day.
Q: Can I filter similarity results using my own metadata?
A: Yes. Advanced Search allows you to combine similarity with filters based on your internal metadata, such as release date, territory, clearance status, or anything you attach as custom tags.
Q: Does Cyanite offer a Prompt-based Search?
A: Yes. Cyanite supports natural language search, enabling users to search for music using descriptive queries. Musiio does not offer Free Text Search.
Q: What are Cyanite’s rate limits for similarity search?
A: Cyanite supports up to 10 search requests per second by default, enabling real-time similarity workflows for user-facing discovery features.
Q: How is Cyanite priced for teams migrating from Musiio?
A: API access typically includes a base fee. Search usage and advanced features are volume-based. For larger volumes and enterprise use cases, bulk discounts are available.