Reflex API
Morph: Reflex is a specialist model that classifies every turn in an agent loop, returning one deterministic answer from predefined options without generating text.
Morph: Reflex is a small, very fast specialist model built by Morph to sit alongside larger coding models inside an agent. Instead of producing free text, it reads the agent's state once and answers every question from that single read, labeling each turn in the loop. Because its output is constrained to predefined options, answers cannot fall outside the choices provided.
The model is designed for classification tasks inside agentic workflows: labeling or categorizing turns, detecting when an agent loops, detecting when a user gives up, and detecting when a jailbreak attempt slips past. Its answers are deterministic, which makes the behavior predictable for routing and monitoring logic.
Reflex fits workflows where a quick, constrained decision is more useful than a written response. If no existing reflex covers a particular need, Morph lets you create one.
Updated:
MODEL ID
reflex- Context
- 65.536K
- Max output
- 58.982K
- Input
- text / image
- Output
- decisions
Highlights
- Labels every turn in an agent loop
- Deterministic answers, no free text
- Answers constrained to predefined options
- Reads state once per request
- Small and very fast specialist model
- Create a reflex if none exists
Providers
Compare API prices across providers. Select a provider to view all charges and conditions.
| $0.042 | $0 | $0.01 |
Decisions Quick Start
Send state and narrow typed questions to /api/v1/decisions and read one answer per question: noul returns a probability from 0 to 1, choice the selected option with its distribution, and score a numeric score for ordered criteria. This endpoint is separate from the Chat Completions, Messages, and Responses surfaces.
curl https://openapi.felo.ai/api/v1/decisions \
-H "Authorization: Bearer $FELO_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "reflex",
"state": "Help! My payouts have been failing for 3 days.",
"questions": {
"is_urgent": {
"type": "noul",
"instructions": "Does this message convey urgency?",
"criteria": {"true": "Time-sensitive", "false": "No urgency"}
},
"department": {
"type": "choice",
"instructions": "Which team should handle this?",
"criteria": {
"billing": "Payments, invoicing, refunds",
"technical": "Bugs, outages, integrations",
"sales": "Pricing, upgrades, new accounts"
}
},
"frustration": {
"type": "score",
"instructions": "How frustrated is the customer?",
"criteria": ["Calm", "Frustrated", "Very angry"]
}
}
}'Frequently asked questions
What is Reflex?
Morph: Reflex is a small, very fast specialist model built by Morph to sit alongside larger coding models inside an agent. Instead of producing free text, it reads the agent's state once and answers every question from that single read, labeling each turn in the loop. Because its output is constrained to predefined options, answers cannot fall outside the choices provided. The model is designed for classification tasks inside agentic workflows: labeling or categorizing turns, detecting when an agent loops, detecting when a user gives up, and detecting when a jailbreak attempt slips past. Its answers are deterministic, which makes the behavior predictable for routing and monitoring logic. Reflex fits workflows where a quick, constrained decision is more useful than a written response. If no existing reflex covers a particular need, Morph lets you create one.
How is usage billed?
Your bill depends on the provider you use, its prices and your usage. Longer inputs or other conditions may have different prices.
What does Morph: Reflex do?
It handles classification tasks and labels every turn in an agent loop. It reads the state once and answers every question from that single read.
Does Morph: Reflex generate text?
No. It generates no text, so answers cannot fall outside the options provided.
What tasks is Morph: Reflex suited for?
It is suited for labeling or categorizing turns in an agentic workflow, detecting when an agent loops, detecting when a user gives up, and detecting when a jailbreak attempt slips past.
How does Morph: Reflex differ from larger coding models?
Morph builds small, very fast specialist models designed to sit alongside larger coding models inside an agent. Reflex provides deterministic answers constrained to predefined options rather than free text.
Can I add a reflex for a task that isn't covered?
Yes. If there is no reflex for what you need, you can create one.