# gligabao

- [gligabao](https://gligabao.mintlify.app/index.md): Minimal memory system for AI. A small, self-hosted HTTP service that turns text into a temporal knowledge graph and answers questions from it.
- [Quickstart](https://gligabao.mintlify.app/getting-started/quickstart.md): Build gligabao, start the server, then remember and query your first facts.
- [How it works](https://gligabao.mintlify.app/getting-started/how-it-works.md): Remember, query, and the bi-temporal knowledge graph underneath.
- [Configuration](https://gligabao.mintlify.app/getting-started/configuration.md): The gligabao.toml config file, API keys, vector store, prices and eval/admin modes.
- [Database](https://gligabao.mintlify.app/getting-started/database.md): Choose SQLite or PostgreSQL for the main database.
- [Datasets](https://gligabao.mintlify.app/concepts/datasets.md): A dataset is the hard, per-user container for everything gligabao stores.
- [Artifacts](https://gligabao.mintlify.app/concepts/artifacts.md): Artifacts are the raw inputs gligabao ingests, and the source every entity and fact is traced back to.
- [Temporal facts](https://gligabao.mintlify.app/concepts/temporal.md): How gligabao tracks when facts were true and when it learned them, following Graphiti's bi-temporal model.
- [Identity](https://gligabao.mintlify.app/concepts/identity.md): Every request to gligabao says which user it acts for, with a user_id.
- [Sessions](https://gligabao.mintlify.app/concepts/sessions.md): Label one user or agent run and trace every artifact, entity and fact back to it.
- [Chunking](https://gligabao.mintlify.app/guides/chunking.md): How long texts are split into overlapping chunks that are extracted and embedded individually.
- [Images](https://gligabao.mintlify.app/guides/images.md): Read text and classify the theme of an image with a vision LLM.
- [Timeseries](https://gligabao.mintlify.app/guides/timeseries.md): Attach values over time, such as prices, to the entities and facts gligabao remembers.
- [Web search](https://gligabao.mintlify.app/guides/web-search.md): Enrich what is remembered with the latest world events about key entities.
- [Geo](https://gligabao.mintlify.app/guides/geo.md): Find the places among the entities of a text and store their coordinates.
- [Vector storage](https://gligabao.mintlify.app/guides/vectors.md): Where embeddings are stored: chromem-go (embedded) or Pinecone (hosted).
- [Evals and ATIF trajectories](https://gligabao.mintlify.app/evals/evals.md): End-to-end evals that run remember and query against a real LLM, recorded as ATIF trajectories.
- [Eval experiments](https://gligabao.mintlify.app/evals/experiments.md): Named series of eval runs for comparing models and settings.
- [Evals of trained models](https://gligabao.mintlify.app/evals/training.md): The Eval button of a trained model: it is asked the cases of an eval, questions it never saw, one per prompt, and an LLM judges the answers in batches. Why one question per prompt matters for Gemma 4 4B, 12B and 31B.
- [Training](https://gligabao.mintlify.app/training/overview.md): How gligabao fine-tunes Gemma-4 on a cloud GPU: the pieces, how a job flows from the training API to the provider, and where each part is documented.
- [Running on Modal](https://gligabao.mintlify.app/training/running.md): The ai_train code layout, the modal commands, the sidecar settings and the local tests.
- [Modal app lifecycle](https://gligabao.mintlify.app/training/modal.md): List the active Modal apps of this system; deploy, watch, cancel and stop each one, and list their volumes, from the sidecar and the admin UI, without the modal CLI.
- [Training jobs](https://gligabao.mintlify.app/training/jobs.md): The TrainingJob struct, the training endpoints that start and list jobs and stream their progress, the gb_training_job_runs and gb_training_job_events tables and the Go client calls.
- [Training datasets](https://gligabao.mintlify.app/training/datasets.md): The dataset file format, the training datasets directory and its dataset.toml manifests, how the Go server lists training datasets, and how a dataset reaches the sidecar inside a job start.
- [Model sizes](https://gligabao.mintlify.app/training/models.md): Gemma-4 4B, 12B and 31B: base models, training backends, GPUs, LoRA settings, cost estimates and troubleshooting.
- [Trained models](https://gligabao.mintlify.app/training/trained-models.md): Where a job's weights are saved, how a completed job becomes a gb_trained_models row, how inference picks a model, and how to add an inference endpoint.
- [Model deploying](https://gligabao.mintlify.app/training/deploying.md): How a trained model is handed to your code: the deployer hook, how gligabao fetches and verifies the weights, the gb_trained_model_deployments record, the API and the limits.
- [Training data audit](https://gligabao.mintlify.app/training/audit.md): The Audit button of a trained model: it is asked its own fine-tuning dataset's questions and an LLM judges the answers, to check that the training took and to find bad training data. Not an eval: it says nothing about generalization.
- [Sidecar training API](https://gligabao.mintlify.app/training/sidecar-api.md): The REST endpoints of the sidecar web API: training jobs and progress, inference and eval examples. There are no dataset endpoints.
- [Providers and extending](https://gligabao.mintlify.app/training/providers.md): How the provider abstraction keeps the training code cloud-neutral, how to add a provider, and how to add a dependency, a model size or a function.
- [Remember](https://gligabao.mintlify.app/api-reference/remember.md): Extract entities and temporal facts from text or an image and store them in the graph.
- [Query](https://gligabao.mintlify.app/api-reference/query.md): Answer a natural-language question from the knowledge graph, with the evidence used.
- [Graph export](https://gligabao.mintlify.app/api-reference/graph.md): Export the whole knowledge graph of a dataset as JSON.
- [Ontologies](https://gligabao.mintlify.app/api-reference/ontologies.md): Load OWL ontologies describing the concepts of a domain and how they relate.
- [Timeseries](https://gligabao.mintlify.app/api-reference/timeseries.md): Store timeseries and anchor them to artifacts, entities and facts.
- [Endpoints overview](https://gligabao.mintlify.app/api-reference/overview.md): Every HTTP endpoint gligabao exposes.
- [Embedding in Go programs](https://gligabao.mintlify.app/embedding/go-library.md): Use gligabao as a Go library through the root package.
- [Hooks](https://gligabao.mintlify.app/embedding/hooks.md): Plug your own code into the gligabao server when embedding it in a Go program.
- [Admin UI extensions](https://gligabao.mintlify.app/embedding/admin-extensions.md): Add your own views, buttons, modals and blocks to the gligabao admin UI from a Go program, with plain JavaScript and no build step.
- [Remember manager](https://gligabao.mintlify.app/embedding/remember-manager.md): How the remember pipeline is orchestrated internally.
- [Inferences](https://gligabao.mintlify.app/operations/inferences.md): Every model invocation is logged with its model, tokens and cost.
- [Python sidecar](https://gligabao.mintlify.app/operations/sidecar.md): The FastAPI service the Go server talks to, and that proxies every request to its backend (Modal): its routes, backends, settings, container and Go client.
- [PyPI packages](https://gligabao.mintlify.app/operations/pypi.md): Installing gligabao from pip: the client (gligabao) and the server (gligabao-server), and how run_pypi.sh builds and uploads them.
- [Building and versioning](https://gligabao.mintlify.app/operations/versioning.md): Static builds, the VERSION file, the changelog and the version endpoint.
- [Not included](https://gligabao.mintlify.app/appendix/not-included.md): Things gligabao deliberately does not do, or does not do yet, with the reasoning.
- [Timeseries design proposal](https://gligabao.mintlify.app/appendix/timeseries-design.md): Design proposal for timeseries support in remember.

## Optional

- [GitHub](https://github.com/deepblue-phoenix/gligabao)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.