Deep research needs different compute than writing a subject line. We match every agent to the model best suited to its task — so you get better outputs, faster, without thinking about which model to use.
A market sizing analysis and a subject line A/B test are not the same cognitive task. Routing them to the same model wastes compute on simple tasks and under-resources complex ones.
Execution agents that produce copy variants run on Haiku and return in seconds. Research agents take longer because they're doing more — and should.
When Anthropic ships a stronger model, agents are updated to use it. You never need to manage model versions or prompt compatibility yourself.
The model stack
Bring your own key (BYOK)
If you're on the Starter plan, you bring your own API keys from Anthropic, OpenAI, or Google. Your keys are encrypted with AES-256-GCM before storage and decrypted in memory only at request time.
Growth and above plans use ameena.ai-hosted models — no API key required. We handle the infrastructure, rate limits, and model versioning.
How we secure your API keys →Hosted models (Growth +)
On Growth and CMO Suite plans, we handle the model infrastructure. No API keys, no rate limit management, no version compatibility issues. You focus on the output; we manage the plumbing.
Model updates
When Anthropic ships a better model, we update agent assignments. You get improved output quality without touching a settings page or rewriting prompts. Every agent always runs on the best available model for its task.
30-day free trial. No API key required on Growth and above.
Start free trial →On Starter? Add your API key here →