# HuggingPlace

> HuggingPlace is an AI observability and evaluation platform: log and trace LLM interactions, evaluate output quality, red-team models, curate datasets, run benchmarks, and manage inference. Official SDKs for Rust, Python, JavaScript, and Go, plus an MCP server.

## Start Here

- [Platform Overview](https://huggingplace.com/docs/overview): How HuggingPlace objects fit together: collections, pipelines, evaluations, training, endpoints, and runtime.
- [Quick Start](https://huggingplace.com/docs/quickstart): Create a collection, configure a pipeline, run an evaluation, then stage an endpoint.
- [LLM Operating Guide](https://huggingplace.com/docs/llm-guide): Machine-readable guidance for agents that need to operate HuggingPlace workflows safely.
- [How to Integrate](https://huggingplace.com/docs/integration-tutorial): End-to-end integration tutorial: SDK and API paths, tracing, and what each step unlocks.
- [DPO Management Overview](https://huggingplace.com/docs/dpo-overview): Short overview of preference stores and non-goals (no dedicated /api/dpo yet).
- [Shadow Evaluators Overview](https://huggingplace.com/docs/shadow-evaluators-overview): Short design overview of objects, trigger path, and related strategy docs.
- [Topology Cache Overview](https://huggingplace.com/docs/topology-cache-overview): Short design overview of topology vs GraphRAG and required infrastructure.
- [API Overview](https://huggingplace.com/docs/api-overview): Base URLs, auth headers, rate limits, ingestion styles, shadow and topology pointers.

## SDK Libraries

- [Rust SDK](https://huggingplace.com/docs/rust-sdk): Log and trace LLM interactions from Rust. Send-safe futures, typed errors, full tracing + evaluation surface.
- [Python SDK](https://huggingplace.com/docs/python-sdk): pip install huggingplace-sdk. Async log(), sessions, run/event tracing, and the evaluate family.
- [JavaScript SDK](https://huggingplace.com/docs/javascript-sdk): npm install huggingplace-sdk. ESM client for Node with logging, tracing, and evaluation helpers.
- [Go SDK](https://huggingplace.com/docs/go-sdk): Standard-library-only Go client. Log(), StartSession, run/event tracing, and Evaluate/EvaluateRAG.

## Core Workflows

- [Collections](https://huggingplace.com/docs/collections): Manage versioned data for retrieval, evaluation, and training lineage.
- [Edge Devices](https://huggingplace.com/docs/edge-devices): Heartbeat, command pull, acknowledgements, command results, and local runtime inventory for on-prem execution.
- [Local Models and Cloud Fallback](https://huggingplace.com/docs/local-models-fallback): Local-first routing, model readiness, cloud fallback, and the execution metadata operators should track.
- [Provider and Key Governance](https://huggingplace.com/docs/provider-governance): Provider configs, API keys, virtual keys, budgets, and key-based reporting across execution paths.
- [Security and Compliance](https://huggingplace.com/docs/security-compliance): Review flagged logs, define compliance frameworks, scan text, and keep operator guardrails inside the Rust-first control plane.
- [Red Team Testing](https://huggingplace.com/docs/red-team): Test LLM vulnerabilities with automated scanning, adversarial attack generation, and runtime guardrails powered by deepteam.
- [Remote Platforms](https://huggingplace.com/docs/remote-platforms): Capability-aware remote platform lifecycle control for validate, sync, rotate, revoke, and audit history.
- [Runtime Operations](https://huggingplace.com/docs/runtime-operations): Manage runtime files, runtime batches, execution targets, provider configs, and fallback reasons across local and cloud execution.
- [Observability](https://huggingplace.com/docs/observability): Track usage events, routing decisions, remote platform activity, runtime ops, and governance saturation.
- [Pipelines](https://huggingplace.com/docs/pipelines): Author and operate flow steps for retrieval, generation, evaluation, and runtime routing.
- [Evaluations](https://huggingplace.com/docs/evaluations): Run quality checks with metrics + LLM-as-judge outputs tied to stable dataset versions.
- [Shadow Evaluators](https://huggingplace.com/docs/shadow-evaluators): First-class continuous offline judges: suites, sample %, results, and chat logging contract for implementers.
- [Topology Cache](https://huggingplace.com/docs/topology-cache): Org/team/user/session cache layers and shared vs user SQL for multi-tenant product backends.
- [DPO / Preference Management](https://huggingplace.com/docs/dpo-management): Mint, store, filter, and train on dpo_v1 preference pairs using events, collections, and training APIs.
- [Edge Devices](https://huggingplace.com/docs/edge-devices): Device registration, command dispatch, acknowledgements, result payloads, and heartbeat flows.
- [Local Models with Cloud Fallback](https://huggingplace.com/docs/local-models-fallback): Configure local-first inference and controlled cloud fallback across endpoints and pipelines.
- [Provider and Key Governance](https://huggingplace.com/docs/provider-governance): Provider configs, API keys, virtual keys, budgets, and execution policy controls.
- [Remote Platforms](https://huggingplace.com/docs/remote-platforms): Integrate external control planes for key lifecycle and platform-specific operations.
- [Observability](https://huggingplace.com/docs/observability): Session/run/event telemetry, quality summaries, runtime health, and promotion diagnostics.
- [Private Cloud](https://huggingplace.com/docs/private-cloud): Deploy HuggingPlace and Buttrbase on private infrastructure with environment-safe runtime config.
- [API Reference](https://huggingplace.com/docs/api): Full route reference including chats, evaluate, shadow evaluators, topology cache, sessions, training.

## Tutorials

- [Tutorials](https://huggingplace.com/docs/tutorials): Hands-on workflow drills for operators and deployment teams.
- [Tutorial: Evaluating Output](https://huggingplace.com/docs/evaluation-tutorial): Inline evaluate, RAG, safety, G-Eval, metrics list, and CI gates.
- [Tutorial: Shadow Evaluators](https://huggingplace.com/docs/shadow-tutorial): Configure sample %, suites, chat feed contract, SDK methods, and failure modes.
- [Tutorial: Topology Cache](https://huggingplace.com/docs/topology-tutorial): Layers, shared SQL, teams, warm/invalidate, privacy rules, SDK methods.
- [Tutorial: DPO Management](https://huggingplace.com/docs/dpo-tutorial): dpo_v1 schema, mint channels, filters, training export, Solid ops checklist.

## Reference

- [API reference](https://huggingplace.com/docs/api): every backend route and payload
- [Rust SDK repo](https://github.com/S7-Works/huggingplace-sdks-v2/tree/main/huggingplace-rust-sdk)
- [Python SDK repo](https://github.com/S7-Works/huggingplace-sdks-v2/tree/main/huggingplace-python-sdk)
- [JavaScript SDK repo](https://github.com/S7-Works/huggingplace-sdks-v2/tree/main/huggingplace-sdk)
- [Go SDK repo](https://github.com/S7-Works/huggingplace-sdks-v2/tree/main/huggingplace-go-sdk)
- [MCP server repo](https://github.com/S7-Works/hplace-mcp-rust-v2)
- [Full documentation](https://github.com/S7-Works/huggingplace-docs)

## Guides & Reference

- [API Overview](https://github.com/S7-Works/huggingplace-docs/blob/main/api/overview.md)
- [API Reference](https://github.com/S7-Works/huggingplace-docs/blob/main/api/reference.md)
- [Authentication](https://github.com/S7-Works/huggingplace-docs/blob/main/configuration/auth.md)
- [Backend Environment Reference](https://github.com/S7-Works/huggingplace-docs/blob/main/configuration/backend-env.md)
- [Frontend Environment Reference](https://github.com/S7-Works/huggingplace-docs/blob/main/configuration/frontend-env.md)
- [LLM Providers](https://github.com/S7-Works/huggingplace-docs/blob/main/configuration/llm-providers.md)
- [HuggingPlace: model/adapter shadow eval + promote (accuracy × cost)](https://github.com/S7-Works/huggingplace-docs/blob/main/designs/2026-08-07-model-shadow-promote-lora.md)
- [HuggingPlace: model-shadow workflow as **legos** + schema builder + share](https://github.com/S7-Works/huggingplace-docs/blob/main/designs/2026-08-07-model-shadow-workflow-legos.md)
- [Overview — DPO / preference-pair management](https://github.com/S7-Works/huggingplace-docs/blob/main/designs/2026-08-12-dpo-management-overview.md)
- [Overview — First-class shadow evaluators](https://github.com/S7-Works/huggingplace-docs/blob/main/designs/2026-08-12-shadow-evaluators-overview.md)
- [Overview — Topology cache](https://github.com/S7-Works/huggingplace-docs/blob/main/designs/2026-08-12-topology-cache-overview.md)
- [Getting Started](https://github.com/S7-Works/huggingplace-docs/blob/main/getting-started.md)
- [Backend Installation](https://github.com/S7-Works/huggingplace-docs/blob/main/installation/backend.md)
- [Deployment](https://github.com/S7-Works/huggingplace-docs/blob/main/installation/deployment.md)
- [Frontend Installation](https://github.com/S7-Works/huggingplace-docs/blob/main/installation/frontend.md)
- [Integration Guide](https://github.com/S7-Works/huggingplace-docs/blob/main/integration-guide.md)
- [MCP Server](https://github.com/S7-Works/huggingplace-docs/blob/main/mcp/server.md)
- [How to Integrate HuggingPlace](https://github.com/S7-Works/huggingplace-docs/blob/main/tutorials/00-how-to-integrate.md)
- [Tutorial 1 — Log your first interaction](https://github.com/S7-Works/huggingplace-docs/blob/main/tutorials/01-log-your-first-interaction.md)
- [Tutorial 2 — Sessions and tracing](https://github.com/S7-Works/huggingplace-docs/blob/main/tutorials/02-sessions-and-tracing.md)
- [Tutorial 3 — Evaluating LLM output](https://github.com/S7-Works/huggingplace-docs/blob/main/tutorials/03-evaluation.md)
- [Tutorial 4 — Red-teaming your model](https://github.com/S7-Works/huggingplace-docs/blob/main/tutorials/04-red-teaming.md)
- [Tutorial 5 — Using HuggingPlace from Claude via MCP](https://github.com/S7-Works/huggingplace-docs/blob/main/tutorials/05-mcp-with-claude.md)
- [Tutorial 6 — Configurable shadow evaluators](https://github.com/S7-Works/huggingplace-docs/blob/main/tutorials/06-shadow-evaluators.md)
- [Tutorial 7 — Topology cache (org / team / user + shared SQL)](https://github.com/S7-Works/huggingplace-docs/blob/main/tutorials/07-topology-cache.md)
- [Tutorial 8 — DPO / preference-pair management](https://github.com/S7-Works/huggingplace-docs/blob/main/tutorials/08-dpo-management.md)

## Optional

- [Full docs as one file](https://huggingplace.com/llms-full.txt): every page and guide inlined