About and methodology
How we review.
Multi Agent Navigator is an independent trade magazine on AI agent systems. It combines editorial analysis, a directory with verifiable data and a route to implementation partners.
01
Sources and review date
Every profile lists the sources its statements are based on: documentation, repository, pricing page or the vendor's announcements. Where we checked a profile against these sources, it shows the date of that review.
At the moment 22 of 50 profiles carry a review date. Profiles without one show their sources but no date.
02
Status
- Active
- Under active development and can be recommended.
- Maintenance
- Bug fixes only, no new features. Use with care for new projects.
- Superseded
- The vendor recommends a successor, which we link in the profile.
- Discontinued
- No longer offered or maintained. We no longer recommend it.
03
Ratings from 1 to 5
The key figures in profiles and in the cockpit are editorial assessments on a scale from 1 to 5. They compare systems with each other, not with an ideal.
- Usable without developers
- 1 means code is required, 5 means fully visual.
- Scalability
- How well the system handles load, teams and production use.
- Integrations
- Range of ready-made connections to models, data and tools.
- Versioning
- Whether agents, prompts and flows can be versioned and rolled back.
- Observability
- Tracing, logs and evaluation of agent runs.
04
How the site is funded
- Partner inquiries: partners pay only for an inquiry they accept. There is no profile fee, and being a partner never changes a rating. How it works
- Affiliate links: some links to vendors earn a commission. They are marked and do not influence the content.
Found an error?
Tell us. We correct facts with a source and update the review date.
Contact the editors