Per-agent control
Support can run a stronger model while a simple router stays light and fast.
How it works
Match the model to the work
- 01
Create the agent
Start with the job — receptionist, support, lead qualifier, or a focused subagent.
- 02
Choose a model — or your key
Use Embage models billed in AI credits, or paste an OpenRouter key so chat is billed to you instead.
- 03
Tune as you learn
If answers are thin, step up. If costs climb on simple FAQs, step down.
- 04
Mix across helpers
The main agent and its subagents can use different models for different work.
Capabilities
Why model choice belongs in a no-code builder
You should not need an engineering team to change brains.
Per-helper control
A knowledge search helper and a summarizer do not have to share the same model.
Bring your own OpenRouter key
Paste an OpenRouter key in the builder. Chat runs on your key, so Embage does not charge AI credits for those sessions. OpenRouter is the BYOK provider today.
Cost that matches usage
Stay on Embage AI credits, or skip them with your own key. Knowledge, datastore, and integration quotas still follow your plan.
Quality where it counts
Use a stronger model for long investigations, refunds, or ambiguous questions.
Swap without rebuilds
Change the model in settings — keep knowledge, datastores, and tools as they are.
Fits the no-code flow
Model choice is a field in the builder, not a deployment pipeline.
In practice
Simple rule of thumb
Start practical, then adjust from real conversations.
Receptionist & routing
Prefer faster, lighter models when the job is greet, classify, and hand off.
Deep support
Prefer stronger models when policy nuance and multi-step troubleshooting matter.
Structured extraction
Post-conversation reviewers often want a model that follows field shapes carefully.
High-volume FAQs
If most chats are short and knowledge-backed, keep the main model lean.
FAQ
Common questions
Do I lock into one model forever?
No. You can change models as your traffic, quality bar, or budget changes — without rebuilding the agent.
Can subagents use different models?
Yes. That is one of the main reasons to split work into helpers — each job gets the brain it deserves.
Can I bring my own API key?
Yes. OpenRouter is supported today. Paste your key when you create or edit an agent. Chat then uses your OpenRouter account, and Embage does not charge AI credits for those sessions. Voice is still priced per session. Knowledge searches, datastore writes, and integration calls still count against your plan quotas.
How does this relate to pricing?
Without a key, chat uses Embage AI credits. With your OpenRouter key, those credits are skipped. Either way you can still pick the model that fits each agent or helper.