# Overmind - [Overmind overview](https://docs.overmindlab.ai/index.md): Understand your AI application, prepare data, evaluate changes and train models from your coding agent. - [Connect your coding agent](https://docs.overmindlab.ai/how-to/connect.md): Connect MCP and local file transfer to the same account, project and deployment. - [Tutorials](https://docs.overmindlab.ai/tutorials/index.md): Learn dataset-to-model workflows, tracing and evaluation through your coding agent. - [How to train a decision model](https://docs.overmindlab.ai/tutorials/first-dataset.md): Prepare data, train a decision model and benchmark its results with one prompt. - [How to fine-tune a language model](https://docs.overmindlab.ai/tutorials/fine-tune-language-model.md): Prepare examples, fine-tune a language model and compare its responses with the base model. - [Record your first trace with a prompt](https://docs.overmindlab.ai/tutorials/first-trace.md): Let your coding agent write, run and inspect a tiny Python application. - [Run your first evaluation](https://docs.overmindlab.ai/tutorials/first-evaluation.md): Ask your agent to run a small check and explain the result. - [Map an agent repository](https://docs.overmindlab.ai/how-to/map-agent.md): Discover capabilities, review the source-grounded map and sync it to a project. - [Instrument an application](https://docs.overmindlab.ai/how-to/instrument.md): Add trace boundaries, capability identity and useful input/output evidence. - [Import traces from another provider](https://docs.overmindlab.ai/how-to/import-traces.md): Inspect source boundaries, confirm capability mapping and import retained traces. - [Investigate a failure](https://docs.overmindlab.ai/how-to/investigate.md): Follow a score, error or latency signal back to recorded evidence. - [Prepare data for a task](https://docs.overmindlab.ai/how-to/prepare-data.md): Retain original sources, author a reproducible transformation and verify the published output. - [Separate training and held-out data](https://docs.overmindlab.ai/how-to/partition-data.md): Save group-preserving train, development, calibration and final roles. - [Run an evaluation](https://docs.overmindlab.ai/how-to/evaluate.md): Freeze a dataset and checks, compare targets and inspect paired results. - [Compare decision models](https://docs.overmindlab.ai/how-to/compare-decisions.md): Compare foundations and trained probability models on frozen suites. - [How to train a model](https://docs.overmindlab.ai/how-to/train.md): Fine-tune a language or decision model, follow training and assess the result. - [Monitor training and inspect checkpoints](https://docs.overmindlab.ai/how-to/monitor-training.md): Read saved progress, development checks and failed stages without starting more work. - [Improve prompts, code or model choice](https://docs.overmindlab.ai/how-to/optimise.md): Run a bounded repository experiment against frozen data and checks. - [Serve and activate a model](https://docs.overmindlab.ai/how-to/serve.md): Verify a trained deployment and connect an application to its supported inference contract. - [Manage project access and API keys](https://docs.overmindlab.ai/how-to/manage-access.md): Choose credential scope, issue keys and keep account and project boundaries explicit. - [Recover interrupted work](https://docs.overmindlab.ai/how-to/recover.md): Resume from receipts without duplicating uploads, transformations or provider calls. - [Self-host Overmind](https://docs.overmindlab.ai/how-to/self-host.md): Configure the local Compose deployment and its external provider dependencies. - [MCP interface](https://docs.overmindlab.ai/reference/mcp.md): Authentication, discovery, tool families, resources and asynchronous receipts. - [Fraud example dataset](https://docs.overmindlab.ai/reference/example-dataset.md): The ecommerce transaction source used in the walkthroughs. - [CLI commands](https://docs.overmindlab.ai/reference/cli.md): Account connection, source transfer, packages, repository sync and optimiser commands. - [Python tracing SDK](https://docs.overmindlab.ai/reference/python-sdk.md): Initialization, capture controls, span decorators and evaluation declarations. - [OTLP traces and attributes](https://docs.overmindlab.ai/reference/otlp.md): The trace ingest protocol and identity fields. - [Datasets, sources and consumer formats](https://docs.overmindlab.ai/reference/datasets.md): Identities, file limits, query bounds and training/evaluation row shapes. - [Retained transformation packages](https://docs.overmindlab.ai/reference/pipelines.md): Revision identity, execution, branching, validation and publication contracts. - [Evaluation matrix and runs](https://docs.overmindlab.ai/reference/evaluation.md): Row definitions, deterministic span fields, targets and result interpretation. - [Training inputs, monitoring and receipts](https://docs.overmindlab.ai/reference/training.md): Exact preparation, frozen recipes, decision qualification and measured progress. - [Inference contracts](https://docs.overmindlab.ai/reference/inference.md): Chat completions, typed decisions, durable requests and deployment state. - [REST API map](https://docs.overmindlab.ai/reference/api.md): Authentication, schema discovery and the current resource families. - [Deployment configuration](https://docs.overmindlab.ai/reference/configuration.md): Server settings and operating boundaries for self-hosted Overmind. - [Glossary](https://docs.overmindlab.ai/reference/glossary.md): Terms used in the Console, MCP, CLI and result receipts. - [Licensing](https://docs.overmindlab.ai/reference/licensing.md): AGPL-3.0 for the platform, MIT for the SDK and CLI, and a commercial licence for different terms. - [Working with a native coding agent](https://docs.overmindlab.ai/explanation/native-agent.md): How responsibilities are shared between your agent, CLI, MCP and Console. - [Capabilities and repository evidence](https://docs.overmindlab.ai/explanation/capabilities.md): How a source map relates to observed behaviour without replacing it. - [Data lineage and held-out evidence](https://docs.overmindlab.ai/explanation/data-lineage.md): Why original sources, immutable cells and group identity matter. - [How evaluation works](https://docs.overmindlab.ai/explanation/evaluation.md): The relationship between live scores, frozen comparisons and decision benchmarks. - [Training, benchmarking and serving](https://docs.overmindlab.ai/explanation/model-lifecycle.md): Why a trained artifact, a ready deployment and a live route are different states. - [Decision models and Jev](https://docs.overmindlab.ai/explanation/decision-models.md): Distinguish typed probability prediction, task training and expectation scoring. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.