OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Search & Research AI

Verifai

VerifAI is an open-source document question-answering and generative search engine that retrieves relevant files, produces referenced answers and checks generated claims against their sources with a separate model.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Verifai — OpenViST project profile in Search & Research AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use VerifAI when you need question answering over a document collection and want generated responses linked to retrieved sources with a separate verification stage.

Who it is for

  • Researchers working with biomedical or other document collections
  • Teams that want question-led search over local files
  • Developers configuring a document-answering system with selectable embedding and language models

Documented uses

  • Index supported local documents for searching
  • Ask questions and retrieve relevant source documents
  • Generate answers accompanied by references
  • Check generated claims against source documents and flag possible hallucinations or misinterpretations
Before you choose

Requirements & limits

The project is identified as open source under the AGPL-3.0 license. Its README points to a deployed app and documents a self-hosted setup.

  • The described workflow depends on indexing source documents rather than searching the open web.
  • Self-hosting involves PostgreSQL, OpenSearch and Qdrant, with Docker required for the documented datastore setup.
  • The README describes configurable Hugging Face embedding models and OpenAI API-standard language-model endpoints, including vLLM, NVIDIA NIM, Ollama, OpenAI and Azure.
  • The repository describes a second model for checking generated text against sources; this is a verification step, not a guarantee that answers are error-free.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Prepare a self-hosted VerifAI installation

Install the documented datastore components and configure the core service before indexing a document collection.

Before you start

  • The VerifAI repository or a downloaded release
  • Python and a virtual environment
  • Docker with its daemon running
  • PostgreSQL, OpenSearch and Qdrant components set up through the project script
  • A configured embedding model and language-model endpoint with the required key

Steps

  1. Clone the repository or download a release, then create and activate a Python virtual environment.

  2. Install the backend requirements with pip install -r backend/requirements.txt.

  3. Configure backend/.env from .env.local.example with database, datastore, model endpoint, embedding and verification settings.

  4. Run the project datastore installation script with Docker available.

Expected outcome

A configured core installation with its documented datastore components prepared for the next stage, indexing files.

Guide limitations

  • The README says PostgreSQL development headers may be needed if psycopg2 installation fails.
  • On a clean instance, the README describes additional PostgreSQL repository setup.
  • The guide covers preparation only; it does not establish that a particular environment will install successfully.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
15 October 2026
Original website title
nikolamilosevic86/verifAI · Official homepage title · retained in its source language
How we review sourcesSuggest a correction