What is the Data and Search Platform?

The Data and Search Platform is Aleph Alpha’s ingestion and retrieval platform. It extracts, chunks, and embeds your documents, and serves search results with source citations to your AI applications. This article describes what the platform does, when to use it, and how it differs from the previous generation of the platform.


What the Data and Search Platform does

Document search requires several components: structure extraction from PDFs and Office documents, chunking that preserves context, embedding generation and storage, and a way to combine keyword and semantic search. All of them must keep working as document volume grows.

The Data and Search Platform provides these components as one pipeline:

Documents -> Ingestion -> Embeddings -> Retrieval -> Search results

You upload documents, run an ingestion workflow on them, and query the result over HTTP. Hybrid search (dense vectors plus keyword/BM25) and per-request metadata filtering are included.

When to use it

Use the Data and Search Platform from initial experimentation through production. The same search store and workflow configuration applies at every stage. As usage grows, you change only the parameters and the infrastructure sizing.

How this differs from the previous generation

The built-in presets cover document ingestion and retrieval: they convert files into searchable chunks. Workflows themselves are a general pipeline that users can extend (see For workflow authors). The goal is for users to write the workflows their integration requires for other data processing tasks as well. If the built-in presets do not cover your use case, ask your data team to build a workflow for it.

Compared with the previous generation, the object model is simpler:

  • One HTTP API surface instead of two separate services (previously pharia-data-api for ingestion and document-index for search).

  • One end-user object, the Search Store, instead of separately managing namespaces, collections, and indexes — see Core concepts.

  • Workflow-based ingestion with built-in presets, instead of manually wiring stages, transformations, and repositories — see Workflows and presets.

Next steps