Ultineo Reasoning
Chain AI reasoning steps, and plug them in anywhere
Ultineo Reasoning turns a task into a reasoning step: an application brings together a model, instructions and the fields to fill in. You chain these applications, each step passes its answer to the next, and the API runs it all from any software.
What Ultineo Reasoning does
- One reasoning step per task
An application does one task: it takes a few fields, follows its instructions and returns a structured answer, on the model you chose for it.
- Steps that follow one another
A chain puts several applications in sequence. The answer of one step becomes the input of the next, and conditions decide the path.
- Built into where you work
The API runs your reasoning from your software, your website or your internal tools, with one key per project.
An example chain: handling a customer request
Four blocks, one launch. Each application works on the model that suits it and passes its answer to the next step.
-
1Trigger
A request comes in
Your software sends the customer’s message through the API.
-
2Application
Summarise the request
Reads the message and extracts the essentials in three lines.
Qwen3.5 · 9B
-
3Condition
Does the answer contain “urgent”?
True: the chain goes on. False: it stops here.
-
4Application
Draft the reply
Writes a reply ready for review, based on the summary.
Mistral Small 3.2 · 24B
How a reasoning step is built
- A model
Each application runs on an open model picked from the library: compact and fast, or larger and more thorough.
- Instructions
You write what the application must do and the shape of its answer. The instructions stay yours: those who run the application never see them.
- Fields
The application asks for the few fields it needs, text or an image depending on the model. Filling them in is enough to run it.
An application and a run, in plain view
On the left, what you define once. On the right, what you get at each launch.
The application
Triage a customer request
- Model
- Qwen3 · 4B
- InstructionsVisible to you only
- Classify the message as low, normal or urgent.
- Fields
- Customer messageLong text, required
A runFinished
- Customer message
- My parcel has not arrived and the tracking has not moved for a week.
- Answer
- urgent
Open models, one per step
- A library of models
Qwen, Gemma, Mistral, Granite, Phi, DeepSeek and other families of open models, in several sizes.
- The right model for each task
Text, image, code, translation or step-by-step reasoning: each application keeps its own model, which you can change at any time.
- No provider key
You pick the model; Ultineo runs it and measures the usage. There is no provider key to manage.
Run your reasoning from any software
- One key per project
A project key authenticates every call and reaches only the data of its own project.
- One call, one run
One call runs an application with its fields; another reads the run back until its answer is there.
- Anywhere an HTTPS request can start
Business software, website, back office, script or automation tool: the API exchanges JSON over HTTPS.
A catalogue of ready-made applications
Don’t want to start from a blank page? The catalogue brings together applications already written for 40 trades, from healthcare and law to construction and events.
- Visit Note Structurer
Turns your shorthand consultation jottings into a clean SOAP-format visit note ready for the record.
Ministral 3 · 14B
- Matter Intake Router
Returns the practice area, urgency level and routing recommendation for a new client enquiry.
Ministral 3 · 14B
- Listing Copy Builder
Returns a ready-to-publish portal listing description built only from the property facts you supply.
Ministral 3 · 14B
- Menu Description Writer
Returns an appetising menu description for a dish, written in your venue's register.
Ministral 3 · 14B
- Product Listing Writer
Returns a ready-to-publish title, hook and description for one product.
Ministral 3 · 14B
- Job Ad Composer
Returns a ready-to-post job advert built from your role brief.
Ministral 3 · 14B
How to get started
- Tell us what you need
Write to us and choose Reasoning: the tasks to hand over, the software to connect and the expected volume.
- We open your project
Our team opens your project on the plan that fits your usage.
- Build and measure
You create your applications, chain them, and each run is counted against your plan’s allowance.
Put your reasoning to work
Tell us which tasks to chain and we will open your access.
Questions about Ultineo Reasoning
A run is one execution of an application: the fields you provide, the application’s instructions and the model produce an answer, kept with its duration and the tokens used.
A chain links several applications from a trigger. Each step passes its answer to the next, and conditions test the previous answer to decide the path.
Not to create an application or draw a chain: it is all done in your workspace. The API is for teams that want to run their reasoning from their own software.
Yes. Each application has its own model, picked from the library, and you can change it without touching anything else.
Not yet. There is no self-service sign-up: write to us, choose Reasoning, and our team opens your access.
By plan. Each plan includes a monthly usage allowance measured in tokens; API access is included in paid plans.