Skip to content

DocBot Overview

DocBot is an EBI-wide chatbot designed to provide streamlined access to user documentation across projects. Powered by large language models (LLMs), DocBot assists users by answering queries based on the specific documentation available for a given project. Its responses are generated solely from the documentation added to its knowledge base, and each answer includes source references, allowing users to directly consult the relevant sections. DocBot is built using a Retrieval-Augmented Generation (RAG) approach, leveraging a vector database populated with embeddings derived from documentation provided by various teams.

Screenshot 2025-11-25 at 14.13.22.png

Features

Users can interact with DocBot to retrieve information either from the entire documentation set or from a specific knowledge base which can be done by changing the project dropdown.

Screenshot 2025-11-25 at 14.14.08.png

One of DocBot’s key features is conversational context — it retains information from previous messages within a session to provide more detailed and context-aware responses. All conversations are stored locally in the browser's local storage, ensuring that they persist even if the browser tab is accidentally closed. To start a new conversation, users can select a project from the Project Dropdown or leave it on All Services. Doing so will clear the previous chat history from local storage and begin a fresh session.

For every response, DocBot includes source links to the relevant documentation. This allows users to easily refer to the original content for more detailed information or clarification, ensuring transparency and enabling deeper exploration of the topic.

Screenshot 2025-11-25 at 14.15.39.png

Thanks to its conversational awareness, DocBot can handle queries that span documentation from multiple teams. When such a scenario arises, DocBot will prompt the user to specify which team’s documentation they are referring to. Based on the user’s input, it then refines its response to provide more accurate and relevant information tailored to the original question.

Users can provide feedback on individual responses by selecting a thumbs up or thumbs down, along with an optional comment explaining their rating.

Screenshot 2025-11-25 at 14.20.10.png

Scope

DocBot is the EMBL-wide AI Assistant designed to answer user questions specifically about supported EMBL (European Molecular Biology Laboratory) services.

Currently, DocBot is trained to answer questions and retrieve information from the following services, Any questions, tools, or EMBL services not explicitly listed above are currently out of scope and will not be answered by DocBot.DocBot is actively being developed and constantly updating. Additional EMBL services will be added to its supported scope soon.

Annotare

DocBot (EBI-Wide AI Assistant)

EBI Search (Metadata Search)

Ensembl (Genome Annotation)

Europe PMC (Free access to life sciences literature)

GWAS Catalog (Catalog of human Genome-Wide Association Studies)

Identifiers (Stable URLs for Life Science Domains)

IMPC (Mouse Phenotyping)

IntAct (Molecular Interactions)

InterPro (Protein families annotation)

Job Dispatcher (Sequence Analysis Tools)

MetaboLights (Metabolomics)

MGnify (Metagenomics)

PCF (Proteomics Core Facility)

Pfam (Protein Sequence Classification)

PRIDE (Proteomics)

Rfam (Non-coding RNA Database)

RNAcentral (Comprehensive non-coding RNA sequence database)

Training

UniProt (Protein Information)

wwPDB Policy & Procedures (Worldwide Protein Data Bank)

MorPhiC (Molecular Phenotypes of Null Alleles in Cells)