Decision API
Ask typed questions about a text, such as which team, how urgent, or yes or no, and get a choice, a score or a probability back instead of prose. One request answers many questions at once, and only input tokens are billed. This endpoint is in alpha.
Trusted by 500+ Companies, from Startups to Industry Leaders
How to get typed decisions from text with Eden AI?

What is a Decision API?
A decision API runs a model built to answer closed questions about a text instead of writing one. You define the possible answers, and the model returns a probability for each of them in a single pass. Typical uses are ticket routing, content triage, lead scoring, intent detection and checking an agent's output. Because the answer comes with its probabilities, you can set a threshold and send uncertain cases to a human.
Replace Prompt-and-Parse With One Typed Call
Send the text as state and name each question with its type and its options. The response gives the selected choice with its confidence and full probability distribution, a score on your scale, and a probability for each yes or no question. There is no free text to parse and no malformed JSON to retry.
In this test, the ticket went to billing with full confidence, frustration scored 0.99 on a 0 to 2 scale, and the churn probability came out at 0.59. The call took about 0.6 seconds and cost $0.000018. See how cheap decision models cut the cost of high-volume AI calls.
.avif)

8 Decision Models, Two Served From the EU
Call every decision model with the same API key and the same request format. Prices are per input token, and output is free. GreenPT Metis and the EU deployment of Microsoft Decision 1 also answer on the Eden AI EU endpoint.
Eden AI lists 8 decision models from TypeSafe, Microsoft, OpenAI, Qwen and GreenPT. They all answer the same three question types: a choice among options, a score on a scale, and a yes or no probability. Run them on your own tickets, reviews or messages and compare their answers before choosing one. To learn how these models differ from chat models, read Jev, a new kind of AI model built for decisions.
Value delivered
Save time
Integrate once and access hundreds of models without managing multiple APIs.
Model updates, provider changes, and new releases are handled transparently.

Reduce costs
Use the most efficient model for each need. Avoid vendor lock-in and adapt quickly to pricing or performance changes.

Reduce risk
Built-in fallback mechanisms protect applications from model outages. Routing flexibility allows rapid adaptation to evolving technologies and providers.

They are using Decision API
Got specific needs? We're here to help
AI API Gateway
5.5% platform fee
Advanced AI Platform
Custom price
Other features

Reorder retrieved documents by relevance to a query before they reach your LLM. The Eden AI Rerank API follows the Cohere v2 format and gives you 16 reranker models from 9 providers behind one API key.

Unlock the power of data privacy with our Data Anonymization API. Effortlessly protect sensitive information while maintaining data integrity for your applications.

Unlock the power of emotion detection with our AI API. Effortlessly analyze and interpret emotions from facial expressions to enhance user experiences and drive engagement.
Start building with Eden AI
A single interface to integrate the best AI technologies into your products.
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)
.png)

.png)