# DecimalAI

## Docs

- [Changelog](https://docs.decimal.ai/changelog.md): Recent changes to the DecimalAI platform, SDK, and GitHub Action.
- [Concepts](https://docs.decimal.ai/concepts.md): Every key concept in DecimalAI — how traces, manifests, evaluations, and multi-agent systems work together.
- [Evaluation](https://docs.decimal.ai/concepts/evaluation.md): Evaluators, eval scores, eval verdicts, and the unified decision engine that combines quality + compatibility.
- [Execution Model](https://docs.decimal.ai/concepts/execution-model.md): How DecimalAI captures and organizes agent activity — traces, spans, LLM calls, sessions.
- [Multi-Agent Systems](https://docs.decimal.ai/concepts/multi-agent.md): Orchestrators, sub-agents, delegation vs handoff, and drift detection.
- [Skills & Data Pipeline](https://docs.decimal.ai/concepts/skills-and-data.md): Skills (SKILL.md), datasets, export formats (SFT/DPO), replay, and repair.
- [Versioning & Compatibility](https://docs.decimal.ai/concepts/versioning.md): Manifests, components, compatibility verdicts, and the engine that classifies traces against agent changes.
- [FAQ](https://docs.decimal.ai/faq.md): Quick answers to common questions about DecimalAI.
- [DecimalAI for Engineers](https://docs.decimal.ai/for-engineers.md): Shipping agents to production safely. The deploy-safety reading path: traces, manifests, regression checks, and CI integration.
- [DecimalAI for Platform Teams](https://docs.decimal.ai/for-platform-teams.md): Operating DecimalAI in production: multi-agent debug, webhooks, teams, RBAC, and security.
- [Glossary](https://docs.decimal.ai/glossary.md): Quick-reference definitions for every term used in DecimalAI.
- [Agents](https://docs.decimal.ai/guides/agents.md): Naming, splitting, renaming, and configuring per-agent evaluation. The Agent is the unit of identity that ties every other DecimalAI concept together.
- [Compatibility Policies](https://docs.decimal.ai/guides/compatibility-policies.md): How DecimalAI classifies traces when your agent changes — keep, repair, replay, or drop — and how to customize the rules.
- [Datasets & Training](https://docs.decimal.ai/guides/datasets.md): Build training datasets from production traces and fine-tune models.
- [Evaluations](https://docs.decimal.ai/guides/evaluations.md): Automated quality checks on agent traces — pre-built evaluators, LLM-as-judge, and the evaluation dashboard.
- [Manifests & Versioning](https://docs.decimal.ai/guides/manifests.md): Automatic agent version tracking with compatibility scoring — detect what changed and what broke.
- [Multi-Agent Systems](https://docs.decimal.ai/guides/multi-agent.md): Instrument multi-agent systems — per-sub-agent datasets, orchestration-aware training data, and independent version tracking.
- [Playground](https://docs.decimal.ai/guides/playground.md): Re-run production traces with modified prompts. Edit skills, compare outputs, and iterate on agent behavior — all with your own API keys.
- [Post-deploy Bisect](https://docs.decimal.ai/guides/post-deploy-bisect.md): A regression slipped past the pre-deploy check. Find the manifest version that broke a metric — without re-running anything.
- [Community Registry](https://docs.decimal.ai/guides/registry.md): Browse, install, fork, and publish skills — with real production effectiveness data, not just star counts.
- [Regression Check (GitHub Action)](https://docs.decimal.ai/guides/regression-check.md): Catch agent regressions before they ship — manifest-aware impact analysis on every PR, no eval cases required.
- [Replay](https://docs.decimal.ai/guides/replay.md): Re-run stale traces through your updated agent and compare results.
- [skillevaluation (eval.yaml)](https://docs.decimal.ai/guides/skillevaluation.md): The open A/B benchmark format + runner for agent skills. Run it locally for free on your own API key; push results to DecimalAI; verified runs power the registry.
- [Skills](https://docs.decimal.ai/guides/skills.md): Reusable agent instructions with observability — from SKILL.md to production analytics.
- [SkillScore](https://docs.decimal.ai/guides/skillscore.md): How DecimalAI scores a skill's real, proven effectiveness — and what's public vs. private to your team.
- [Teams & Workspaces](https://docs.decimal.ai/guides/teams.md): Organize agents, traces, and skills by team with role-based access.
- [Tracing](https://docs.decimal.ai/guides/tracing.md): Capture and inspect agent execution traces across any framework.
- [Troubleshooting](https://docs.decimal.ai/guides/troubleshooting.md): Common issues and how to fix them.
- [Webhooks](https://docs.decimal.ai/guides/webhooks.md): Receive HTTP notifications when manifests change, regressions are detected, or usage crosses thresholds.
- [Why DecimalAI?](https://docs.decimal.ai/guides/why-decimalai.md): How DecimalAI compares to LangSmith, Braintrust, Langfuse — and why manifest-aware agent change management is a different category.
- [Catch agent regressions before they ship](https://docs.decimal.ai/introduction.md): The manifest-aware platform for agent change management — catch regressions before deploy, share proven skills, and keep your training data valid as your agent evolves.
- [Migrating from Other Tools](https://docs.decimal.ai/migrations.md): Side-by-side mapping from LangSmith, Braintrust, Langfuse, and Phoenix to DecimalAI.
- [Pricing](https://docs.decimal.ai/pricing.md): DecimalAI plan tiers — Free, Core, Pro, Enterprise — with quotas, rate limits, and Skills Registry features.
- [Quickstart](https://docs.decimal.ai/quickstart.md): See both demos in 2 minutes on seeded data, then wire DecimalAI to your own agent.
- [Roadmap](https://docs.decimal.ai/roadmap.md): What's in flight and planned on the DecimalAI platform & SDK.
- [CLI](https://docs.decimal.ai/sdk/cli.md): Command-line interface for traces, evaluations, skills, datasets, replay, and more.
- [Python SDK](https://docs.decimal.ai/sdk/python.md): The `decimalai` Python package — installation, tracing, manifests, evaluations, replay, datasets, framework adapters, and skills.
- [Datasets & Training](https://docs.decimal.ai/sdk/python/datasets.md): Pull versioned datasets to JSONL or Parquet, push to HuggingFace Hub, or load directly as a HuggingFace Dataset object.
- [Evaluations](https://docs.decimal.ai/sdk/python/evaluations.md): Push scores onto traces, define custom evaluators with @eval, and import results from DeepEval / LangSmith.
- [Framework Instrumentation](https://docs.decimal.ai/sdk/python/frameworks.md): Auto-capture traces from OpenAI Agents, LangChain, LlamaIndex, CrewAI, AutoGen, or any OpenTelemetry-emitting framework — usually with one line of code.
- [AutoGen / AG2](https://docs.decimal.ai/sdk/python/frameworks/autogen.md): Auto-instrument AutoGen and AG2 agent conversations via OpenTelemetry.
- [CrewAI](https://docs.decimal.ai/sdk/python/frameworks/crewai.md): Auto-instrument CrewAI crews. Tasks, agent conversations, and tool calls captured via OpenTelemetry.
- [LangChain & LangGraph](https://docs.decimal.ai/sdk/python/frameworks/langchain.md): Auto-instrument LangChain chains, LangGraph state machines, and AgentExecutor / ReAct agents.
- [LlamaIndex](https://docs.decimal.ai/sdk/python/frameworks/llamaindex.md): Auto-instrument LlamaIndex query engines, retrievers, and synthesizers.
- [OpenAI Agents SDK](https://docs.decimal.ai/sdk/python/frameworks/openai-agents.md): Auto-instrument the OpenAI Agents SDK. Captures LLM generations, tool calls, handoffs, guardrails, and the full manifest.
- [Generic OpenTelemetry](https://docs.decimal.ai/sdk/python/frameworks/otel.md): Capture traces from any framework that emits OTel spans following the gen_ai.* semantic conventions.
- [Init & Setup](https://docs.decimal.ai/sdk/python/init.md): Install the package, initialize the SDK, and configure environment variables. The starting point for every Python SDK user.
- [Manifests](https://docs.decimal.ai/sdk/python/manifests.md): Explicitly register agent configuration for version tracking when auto-detection isn't enough.
- [Replay & Experiments](https://docs.decimal.ai/sdk/python/replay.md): Re-run historical traces against a new manifest, and run controlled A/B comparisons between variants.
- [Skills](https://docs.decimal.ai/sdk/python/skills.md): Sync skill files from disk to platform, pull missing skills, and route to the right skill at runtime.
- [Tracing](https://docs.decimal.ai/sdk/python/tracing.md): Decorators and helpers to record LLM calls, tool calls, and trace boundaries from Python.
- [Security & Data Handling](https://docs.decimal.ai/security.md): What DecimalAI stores, how it's encrypted, retention policies, and deletion.
- [Skills Observability](https://docs.decimal.ai/tutorials/skills-observability.md): Tutorial: Track skill effectiveness from discovery to production analytics.
- [End-to-End Training Pipeline](https://docs.decimal.ai/tutorials/training-pipeline.md): Tutorial: From production traces to a fine-tuned model in one workflow.
- [The 2-Minute Demo](https://docs.decimal.ai/tutorials/two-minute-demo.md): See the impact report and the ranked skills registry on realistic seeded data — one command each, no agent code, no LLM keys.

## OpenAPI Specs

- [openapi](https://docs.decimal.ai/openapi.json)