# little big brain documentation

> Documentation for little big brain, a search platform for AI applications and agents. Vector search finds facts by meaning. A graph query follows them to the evidence. Each answer names the commit it read, so you can replay it. Start with the quickstart, or map what you know from vector stores in the core concepts.

This index covers the public documentation hosted at docs.littlebigbrain.com. Use it to find the right guide, SDK documentation, or full Markdown context file.

## Full context

- [llms-full.txt](https://docs.littlebigbrain.com/llms-full.txt): Single-file Markdown context containing the full docs content.

## Overview

- [little big brain documentation](https://docs.littlebigbrain.com/): Find facts by meaning, follow graph paths to the evidence, and replay any answer. Start with the quickstart, or map what you know from vector stores.

## Start here

- [Quickstart: find it by meaning, prove it by path](https://docs.littlebigbrain.com/start/quickstart/): Write a few records as JSON, find one by meaning, and follow its links to an answer that names the commit it read. About ten minutes. No RDF or graph experience needed.
- [Build a workflow](https://docs.littlebigbrain.com/start/first-workflow/): Test LBB's values.extract template on two records, save your answers as a dataset, then create the workflow and read what it writes.
- [What is little big brain?](https://docs.littlebigbrain.com/start/overview/): little big brain is a search platform for AI applications. Vector search finds facts by meaning, a graph query follows them to the evidence, and every answer names the commit it read.
- [Core concepts](https://docs.littlebigbrain.com/start/concepts/): The little big brain model for developers who know vector stores. Records, types, links, commits, search by meaning, graph queries, and the RDF words you will meet.
- [Glossary](https://docs.littlebigbrain.com/start/glossary/): Short definitions of the words for workflows, quality and review in little big brain, plus the engine words under Details.
- [Load your data](https://docs.littlebigbrain.com/start/load-data/): Load a database and your documents into one graph as JSON, in four calls. Pick a use case, and every example on the page follows it.
- [How search works](https://docs.littlebigbrain.com/start/platform/): Find a starting fact by meaning, follow graph links to the evidence, and replay the answer at the commit it read. One support question, step by step.
- [Use cases](https://docs.littlebigbrain.com/start/use-cases/): Six agent decisions, each with the records to store, the query that answers it, and what to keep so the decision can be replayed.
- [Vector RAG, graph RAG, SQL and full-text search compared](https://docs.littlebigbrain.com/start/database-comparison/): Compare vector, graph, SQL and full-text retrieval for AI agents on one connected-facts question, and see how little big brain combines search by meaning with graph paths.
- [Authentication & endpoints](https://docs.littlebigbrain.com/start/authentication/): Environments, per-environment API keys, bearer auth, base URLs, and how the hosted MCP OAuth flow differs from a static key.
- [Endpoints & base URLs](https://docs.littlebigbrain.com/start/endpoints/): Use the tenant-plus-stack endpoint copied from the console for every hosted data-plane request.
- [Usage & billing](https://docs.littlebigbrain.com/start/billing/): Current little big brain Cloud pricing starts at €19/$21 per month and €0.45/$0.49 per GiB-month. Understand storage, included queries, payments, and retained plan terms.
- [How to model your app](https://docs.littlebigbrain.com/start/model-your-app/): Start with a graph and a token, then add optional namespaces, delegated reads and shared reference datasets.

## SDKs & libraries

- [Choosing an SDK](https://docs.littlebigbrain.com/sdks/overview/): All the ways to talk to little big brain (TypeScript, Python, MCP, and raw HTTP) and when to use each.
- [TypeScript](https://docs.littlebigbrain.com/sdks/typescript/): The TypeScript client for little big brain. Declare types and links, write records, search by meaning, follow links with SPARQL and replay each answer at its commit.
- [Python — lbb](https://docs.littlebigbrain.com/sdks/python/): The Python client for little big brain. Declare types and links, write records, search by meaning, follow links with SPARQL and replay each answer at its commit. Sync and async, with dict or typed responses.
- [MCP server for AI agents](https://docs.littlebigbrain.com/sdks/mcp/): Give an agent tools to search your graph by meaning, follow links with SPARQL, cite the commit it read, and write checked records. Local stdio or hosted OAuth.
- [HTTP API & SPARQL protocol](https://docs.littlebigbrain.com/sdks/http-api/): The HTTP routes every SDK wraps. Auth, idempotency, errors, writes, SPARQL reads, search by meaning, schema, and the native SPARQL 1.1 Protocol endpoint.
- [Error codes](https://docs.littlebigbrain.com/sdks/errors/): Every error code the data plane sends, what it means, and how to fix the request or when to retry it.

## Use cases & guides

- [Add links with an LLM](https://docs.littlebigbrain.com/guides/add-links-with-an-llm/): Some facts live only in your documents. An LLM reads each section and writes them as links. Your code drops links to unknown records, a dry run checks the rest, then you write them.
- [Build with JSON records](https://docs.littlebigbrain.com/guides/without-rdf/): Declare a schema, write records, follow a chain of links, validate writes, and replay a decision, all from JSON and the SDKs. No RDF knowledge needed.
- [Search by meaning](https://docs.littlebigbrain.com/guides/search-by-meaning/): Declare an embedding on a type of record, let the platform keep its vectors current, and search by meaning with filters by type and by link, or inside a SPARQL query. Each hit is checked against the graph.
- [Answer questions in plain words](https://docs.littlebigbrain.com/guides/answer-questions/): Send a question in plain words. The server describes your graph, picks the kind of question, runs queries in a short loop, reads the rows, and answers with the records it cites.
- [Query with SPARQL and validate writes with SHACL](https://docs.littlebigbrain.com/guides/sparql-and-shacl/): Count, filter and follow links between records with SPARQL. Publish SHACL rules for your records, read the conformance report, and reject writes that break a rule.
- [See how a query ran](https://docs.littlebigbrain.com/guides/query-profiles/): Ask for a query profile to see the time, the reads and the join order behind one SPARQL answer. Read the statistics the planner uses for a graph.
- [Replay answers and history](https://docs.littlebigbrain.com/guides/time-travel-audit/): Store the commit behind an agent's answer, replay its query with as_of_commit_seq, and see when an answer changed. Covers pinned reads, validation state, valid time and retention.
- [Reasoning & inference](https://docs.littlebigbrain.com/guides/reasoning/): Let a query also return the facts that follow from your schema, such as "an active study site is a study site". You choose the rules per query with the entailment option.
- [Load RDF data](https://docs.littlebigbrain.com/guides/load-rdf/): Load Turtle, N-Triples, N-Quads or TriG from another triple store or a public dataset, then query it with SPARQL. Covers formats, parameters, limits, large loads and validation.
- [Check documents against each other (experimental)](https://docs.littlebigbrain.com/guides/check-documents-against-each-other/): Upload an order, a delivery note and an invoice. SHACL rules compare their values, and each difference becomes a review item with the quotes of both documents.
- [Check model quality (experimental)](https://docs.littlebigbrain.com/guides/check-model-quality/): See how well the models LBB runs for a graph do their jobs, correct the judge, test other models, tune search, and calibrate thresholds and workflow gates.
- [Datasets of examples (experimental)](https://docs.littlebigbrain.com/guides/datasets/): Keep examples with the values a correct run writes. Save them from a workflow test, add your own, export them as a template's samples, and delete them.
- [Offer Notion, GitHub and Slack in your product (experimental)](https://docs.littlebigbrain.com/guides/developer-owned-integrations/): Register your own provider apps and let each customer authorize selected resources without creating an LBB account.
- [From documents to a graph (experimental)](https://docs.littlebigbrain.com/guides/documents-to-graph/): Upload PDF files, connect the mentions of one thing across them, fill values from text, and read each link with its box on the page.
- [Compare two versions of a workflow (experimental)](https://docs.littlebigbrain.com/guides/experiments/): Run a workflow as it is and with one change on the same examples, grade both against the right values, and see which is better and by how much.
- [Integrations for your customers (experimental)](https://docs.littlebigbrain.com/guides/integrations-for-your-customers/): Connect each customer's sources to a graph of their own, using your backend and your own Notion, GitHub or Slack app.
- [Integrations (experimental)](https://docs.littlebigbrain.com/guides/integrations/): Sync customer sources into a graph on durable workflows, including opt-in developer-owned Notion, GitHub and Slack apps.
- [Ontology starters (experimental)](https://docs.littlebigbrain.com/guides/ontology-starters/): Start a graph from a versioned base ontology for CRM, documents or work management.
- [Ontology suggestions (experimental)](https://docs.littlebigbrain.com/guides/ontology-suggestions/): Review ontology changes from integrations, agents and people before they reach your graph.
- [Dedicated deployments](https://docs.littlebigbrain.com/guides/pro-deployments/): Enterprise runs Little Big Brain on dedicated infrastructure, hosted by us in Europe or in your AWS account, Google Cloud project or Azure subscription.
- [Run a worker in Python (experimental)](https://docs.littlebigbrain.com/guides/python-worker/): Durable workflows in Python with lbb.workflows, the same rules as the TypeScript worker.
- [Read an RDF record](https://docs.littlebigbrain.com/guides/records/): Full IRIs, exact RDF values, independent claims, and snapshot-pinned record reads across HTTP, SDKs and MCP.
- [Ask a person: review in workflows (experimental)](https://docs.littlebigbrain.com/guides/review-in-workflows/): Send a question from a workflow to a person, and use the answer in the graph and in the workflow.
- [Starter workflows (experimental)](https://docs.littlebigbrain.com/guides/starter-workflows/): Create a workflow from a starter workflow on a class, and the platform runs its work on every new or changed instance.
- [Suggest ontology changes from text (experimental)](https://docs.littlebigbrain.com/guides/suggest-ontology-changes-from-text/): Let the platform read your transcripts or documents and propose classes, properties and relations for review.
- [Test a workflow before it writes (experimental)](https://docs.littlebigbrain.com/guides/test-a-workflow/): Run a values workflow on a few instances, read each value with its quote, answer each one, and save the answers as a dataset. Nothing is written to the graph.
- [Settings history of a workflow (experimental)](https://docs.littlebigbrain.com/guides/workflow-settings-history/): See who changed a workflow's settings and when, compare two versions, and go back to earlier settings. Each run records the settings it ran on.
- [Workflow templates (experimental)](https://docs.littlebigbrain.com/guides/workflow-templates/): Run the code of a starter workflow in your own worker, with LBB's models as activities, and change it.
- [Workflows (experimental)](https://docs.littlebigbrain.com/guides/workflows/): Saved state, durable messages and automatic history rollover, stored in object storage.

## Changelog

- [Changelog](https://docs.littlebigbrain.com/changelog/): Product updates for little big brain Cloud and the database engine.
- [Cloud changelog](https://docs.littlebigbrain.com/changelog/cloud/): Updates to the hosted console, Explorer, accounts, access, and billing.
- [Engine changelog](https://docs.littlebigbrain.com/changelog/engine/): Updates to queries, data import, validation, reasoning, performance, and SDKs.