≋FeedPoolPREVIEW

The data an answer starts from

Less searching.
More substance.

Entities, context and sources in a single request. FeedPool prepares data for your AI agents — from cloud models to local assistants.

Choose access How it works

Platform in development · API-first · for many kinds of models

● ● ●FEEDPOOL / LOOKUPDEMO
Agent requestWhat is the weather in Philadelphia?
◈
PLACE / US · PA

Philadelphia

Entity
24°C

Illustrative value

ContextPennsylvania, USA
Example date6 September 2026
Sourceexample.org · demonstration
↳ One structured result
Prepared for different environmentsCloud AI/Self-hosted/API and MCP¹

Three layers. One clear answer.

Stop assembling context
from scratch every time.

The fact you need, together with what it belongs to and where it came from.

01

Entity

We identify the object unambiguously: a place, a publication, an organisation or an issue.

Philadelphia · place
02

Context and facts

We add the values you need, with time ranges, units and relationships.

Pennsylvania · America/New_York
03

Sources and access

We show provenance, links, licence and access conditions.

API · subscription · checked on
Closer to your region. You will be able to set your region in your account. Local entities will get higher priority, while global data and other regions remain available. A single request will be able to specify a different region.

For those who create data

Your data.
A new path to AI.

Do you publish newspapers, run a weather portal or maintain your own catalogue? Present your entities through web forms or an API integration.

Become a publisher
▤
Publishing houses

Issues, sections, dates and access

↗
☀
Weather services

Observations, forecasts and sources

↗
⊞
Catalogues and organisations

Structured records and relationships

↗

Worth knowing

The essentials, briefly.

Is FeedPool my agent's memory?

FeedPool is designed as a shared catalogue of entities and sources. It does not need the history of your conversations for every new request.

Will this work with local models?

Compact JSON is not tied to a particular model. For simple facts a client will be able to use templates without an LLM. Quality and speed depend on your integration and hardware.

Does a subscription unlock paid resources?

No. Paying for FeedPool covers the service. Paid content from external publishers may require a separate subscription or permission.

What is available today?

The platform is in development. This portal presents the product; connecting payments and account areas depends on those services launching. Demo data is not current weather.