Editorial Standards

Independence

AgentBase operates independently. No AI company pays for placement or ranking position in our directory. Rankings are determined solely by benchmark scores, user review data, feature completeness, and update cadence.

How We Add Agents

Every agent is evaluated against our inclusion criteria before listing:

• The product must be publicly accessible (free tier, free trial, or purchasable)
• It must demonstrably use AI/ML as a core capability
• Basic factual information (pricing, features) must be verifiable
• The product must have been updated within the last 12 months

Ranking Methodology

Overall scores are computed from a weighted combination of:

• Benchmark performance — SWE-bench, GAIA, and other standardized tests (40%)
• User ratings — Aggregated from verified review platforms (30%)
• Feature completeness — API access, memory, multimodal, etc. (20%)
• Update velocity — How actively the product is maintained (10%)

Corrections Policy

We correct factual errors as soon as they are identified. If you notice incorrect pricing, outdated features, or wrong benchmark scores, please email [email protected]. We commit to reviewing and responding within 48 hours.

Transparency

When we receive free access to a product for review purposes, we disclose this on the agent's listing page. This does not influence rankings.