# Tracevell documentation

- [Tracevell](https://tracevell.com/docs/overview.md)
- [Getting Started](https://tracevell.com/docs/getting-started.md)
- [Tracing](https://tracevell.com/docs/tracing/introduction.md)
- [Tracing Structure](https://tracevell.com/docs/tracing/structure/overview.md)
- [Trace Functions](https://tracevell.com/docs/tracing/structure/observe-decorator.md)
- [Trace Parts of Your Code](https://tracevell.com/docs/tracing/structure/manual-span-creation.md)
- [Span Types](https://tracevell.com/docs/tracing/structure/span-types.md)
- [Sessions](https://tracevell.com/docs/tracing/structure/sessions.md)
- [User ID](https://tracevell.com/docs/tracing/structure/user-id.md)
- [Metadata](https://tracevell.com/docs/tracing/structure/metadata.md)
- [Tags](https://tracevell.com/docs/tracing/structure/tags.md)
- [Tracing Images](https://tracevell.com/docs/tracing/structure/images.md)
- [Continuing Traces](https://tracevell.com/docs/tracing/structure/continuing-traces.md)
- [Span Attribute Reference](https://tracevell.com/docs/tracing/structure/span-attribute-reference.md)
- [Flushing Spans and Shutting Down](https://tracevell.com/docs/tracing/structure/flushing-and-shutdown.md)
- [Task extraction for AI agent runs](https://tracevell.com/docs/tracing/task-extraction.md): Tracevell extracts the task from every agent run, shows it in the traces table and transcript view, and makes it queryable with SQL.
- [Agent versioning: track prompt and tool changes](https://tracevell.com/docs/tracing/agent-versions.md): Tracevell detects the agents in your traces, versions their system prompts and tools, and tags every trace with the version that ran it.
- [LLM cost tracking](https://tracevell.com/docs/tracing/cost-tracking.md)
- [Session replay for browser agents](https://tracevell.com/docs/tracing/browser-agent-observability.md)
- [LangGraph Visualization](https://tracevell.com/docs/tracing/langgraph-visualization.md)
- [Troubleshooting](https://tracevell.com/docs/tracing/troubleshooting.md)
- [Troubleshooting non-Tracevell SDKs](https://tracevell.com/docs/tracing/troubleshooting-opentelemetry.md)
- [Signals](https://tracevell.com/docs/signals/introduction.md)
- [Quickstart](https://tracevell.com/docs/signals/quickstart.md)
- [Manage Signals from the CLI](https://tracevell.com/docs/signals/cli.md)
- [Clusters](https://tracevell.com/docs/signals/clusters.md)
- [Alerts](https://tracevell.com/docs/signals/alerts.md)
- [Reports](https://tracevell.com/docs/signals/reports.md)
- [Agent Evaluations in Real Environments with Tracevell](https://tracevell.com/docs/agent-evaluations/introduction.md)
- [Agent Evaluations Quickstart with lmnr-cli](https://tracevell.com/docs/agent-evaluations/quickstart.md)
- [Agent Evaluation Concepts - Tasks, Trials, and Verifiers](https://tracevell.com/docs/agent-evaluations/concepts.md)
- [Run Agent Evaluations with lmnr-cli eval run](https://tracevell.com/docs/agent-evaluations/running-evals.md)
- [The Evals Folder Reference for lmnr.yaml, harbor.yaml, and task.toml](https://tracevell.com/docs/agent-evaluations/evals-folder.md)
- [Build an Agent Evaluation Harness That Matches Production](https://tracevell.com/docs/agent-evaluations/production-harness.md)
- [Capture Agent Trajectories in Every Evaluation Trial](https://tracevell.com/docs/agent-evaluations/trajectories.md)
- [Troubleshooting Agent Evaluations](https://tracevell.com/docs/agent-evaluations/troubleshooting.md)
- [LLM Evaluations](https://tracevell.com/docs/evaluations/introduction.md)
- [LLM evaluations quickstart](https://tracevell.com/docs/evaluations/quickstart.md)
- [Evaluation concepts](https://tracevell.com/docs/evaluations/concepts.md)
- [Compare evaluation runs](https://tracevell.com/docs/evaluations/comparing-runs.md)
- [Custom Rendering for Evaluation Traces](https://tracevell.com/docs/evaluations/custom-rendering.md)
- [Datasets for evaluations](https://tracevell.com/docs/evaluations/datasets.md)
- [Manual evaluation API](https://tracevell.com/docs/evaluations/manual-evaluation.md)
- [Self-hosted evaluations](https://tracevell.com/docs/evaluations/self-hosted.md)
- [AI Agent Debugger: Record, Replay, Evaluate](https://tracevell.com/docs/debugger/introduction.md)
- [Set Up the Tracevell Debugger](https://tracevell.com/docs/debugger/setup.md)
- [How the Debugger Runs](https://tracevell.com/docs/debugger/process.md)
- [How Debugger Caching Works](https://tracevell.com/docs/debugger/caching.md)
- [Introduction to Tracevell datasets](https://tracevell.com/docs/datasets/introduction.md)
- [Adding data to Tracevell datasets](https://tracevell.com/docs/datasets/adding-data.md)
- [Tracevell datasets CLI](https://tracevell.com/docs/datasets/cli.md)
- [Platform](https://tracevell.com/docs/platform.md)
- [Viewing Traces](https://tracevell.com/docs/platform/viewing-traces.md)
- [Table views](https://tracevell.com/docs/platform/table-views.md)
- [Slack integration](https://tracevell.com/docs/platform/integrations.md)
- [Live tracing](https://tracevell.com/docs/platform/live-tracing.md)
- [Tracevell Agent](https://tracevell.com/docs/platform/tracevell-agent.md)
- [MCP Server](https://tracevell.com/docs/platform/mcp.md): Connect MCP clients to Tracevell to query traces with SQL and summarize trace context.
- [SQL Editor](https://tracevell.com/docs/platform/sql-editor.md)
- [CLI](https://tracevell.com/docs/platform/cli.md)
- [Terraform Provider for Tracevell Signals and LLM Profiles](https://tracevell.com/docs/platform/terraform.md): Define Tracevell Signals and LLM profiles as code, review changes in a plan, and apply them to every project from CI.
- [Full text search](https://tracevell.com/docs/platform/search.md)
- [Render Templates for Custom Data Views](https://tracevell.com/docs/platform/render-templates.md)
- [Custom Model Costs for Self-Hosted and Fine-Tuned Models](https://tracevell.com/docs/platform/custom-model-costs.md)
- [Usage Warnings and Hard Limits for Data and Signals](https://tracevell.com/docs/platform/usage-limits.md)
- [PII Redaction for AI Agent Traces](https://tracevell.com/docs/platform/pii-redaction.md)
- [Dashboards: Chart Agent Cost, Tokens, and Latency](https://tracevell.com/docs/platform/dashboards.md)
- [Labeling Queues](https://tracevell.com/docs/queues/quickstart.md)
- [LLM Profiles for Custom LLM Providers and API Keys](https://tracevell.com/docs/platform/llm-profiles.md)
- [Playground](https://tracevell.com/docs/playground.md)
- [Agent and LLM Observability Integrations](https://tracevell.com/docs/integrations.md)
- [Anthropic observability](https://tracevell.com/docs/integrations/anthropic.md): Trace every Anthropic API call in TypeScript and Python: prompts, token usage, cost, and latency, all readable in Tracevell's transcript view.
- [OpenAI observability](https://tracevell.com/docs/integrations/openai.md): Trace every OpenAI API call in TypeScript and Python: prompts, token usage, cost, and latency, all readable in Tracevell's transcript view.
- [Gemini observability](https://tracevell.com/docs/integrations/gemini.md): Trace every Gemini API call made with the Google Gen AI SDK: prompts, token usage, cost, and latency, all readable in Tracevell's transcript view.
- [Cohere observability](https://tracevell.com/docs/integrations/cohere.md): Trace every Cohere chat, embed, and rerank call: prompts, token usage, cost, and latency, all readable in Tracevell's transcript view.
- [LiteLLM observability](https://tracevell.com/docs/integrations/litellm.md): Trace LiteLLM calls to any model it routes to with one integration: prompts, token usage, cost, and latency in Tracevell's transcript view.
- [OpenRouter observability](https://tracevell.com/docs/integrations/openrouter.md): Trace every OpenRouter request with the model that served it, token usage, cost, and latency, all readable in Tracevell's transcript view.
- [TypeSafe AI Jev observability](https://tracevell.com/docs/integrations/typesafe.md): Trace every TypeSafe AI Jev call with its state, questions, typed answers, and token usage, all readable in Tracevell's transcript view.
- [Vercel AI SDK observability](https://tracevell.com/docs/integrations/vercel-ai-sdk.md): Trace Vercel AI SDK calls in Next.js and Node.js: model turns, tool calls, streaming, sub-agents, token usage, and cost.
- [eve observability](https://tracevell.com/docs/integrations/eve.md): Trace eve durable agents in TypeScript: every model turn, tool call, token count, and cost, readable as a conversation in Tracevell.
- [LangChain and LangGraph observability](https://tracevell.com/docs/integrations/langchain.md): Trace LangChain chains and LangGraph agents in TypeScript and Python: every graph node, tool call, and LLM call, with token usage and cost, read as a transcript.
- [deepagents observability](https://tracevell.com/docs/integrations/deepagents.md): Trace deepagents runs in Python: the agent root span, every tool call, and every subagent, readable as a conversation in Tracevell.
- [Mastra observability](https://tracevell.com/docs/integrations/mastra.md): Trace Mastra agents, workflows, and tools in TypeScript: model turns, tool calls, sub-agents, token usage, and cost.
- [Pydantic AI observability](https://tracevell.com/docs/integrations/pydantic-ai.md): Trace Pydantic AI agent runs in Python: every model turn, typed tool call, and sub-agent run, readable as a conversation in Tracevell.
- [Microsoft Agent Framework observability](https://tracevell.com/docs/integrations/microsoft-agent-framework.md): Trace Microsoft Agent Framework runs in Python: every agent, model turn, tool call, and workflow step, readable as a conversation in Tracevell.
- [OpenAI Agents SDK observability](https://tracevell.com/docs/integrations/openai-agents-sdk.md): Trace OpenAI Agents SDK runs in TypeScript and Python: agent turns, tool calls, handoffs, and sub-agents in Tracevell's transcript view.
- [Google ADK observability](https://tracevell.com/docs/integrations/google-adk.md): Trace Google Agent Development Kit (ADK) runs in Python: agent invocations, Gemini calls, and tool executions, readable as a conversation in Tracevell.
- [OpenHands SDK observability](https://tracevell.com/docs/integrations/openhands-sdk.md): Trace OpenHands SDK conversations: every agent step, LLM call, and tool call, grouped per conversation as a Tracevell session.
- [Temporal observability](https://tracevell.com/docs/integrations/temporal.md): Trace AI agents that run as Temporal workflows in TypeScript and Python: one trace across the client, the workflow, and every activity.
- [Claude Code observability](https://tracevell.com/docs/integrations/claude-code.md): Trace every Claude Code session with a plugin: each prompt becomes a trace with its LLM calls and tool calls, grouped as a Tracevell session.
- [Codex observability](https://tracevell.com/docs/integrations/codex.md): Trace every Codex session with a plugin: each prompt becomes a trace with its LLM calls and tool calls, grouped as a Tracevell session.
- [Pi observability](https://tracevell.com/docs/integrations/pi.md): Trace every Pi coding agent session: each prompt becomes a trace with its LLM calls and tool calls, grouped as a Tracevell session.
- [Claude Agent SDK observability](https://tracevell.com/docs/integrations/claude-agent-sdk.md): Trace agents built on the Claude Agent SDK in TypeScript and Python, including subagent runs, tool calls, token usage, and cost.
- [OpenCode observability](https://tracevell.com/docs/integrations/opencode.md): Trace OpenCode coding agent sessions: every model turn and tool call, grouped per session and readable as a conversation in Tracevell.
- [Browser Use observability](https://tracevell.com/docs/integrations/browser-use.md): Trace Browser Use agents and watch the browser session recording synced to the agent's steps, LLM calls, and screenshots.
- [Stagehand observability](https://tracevell.com/docs/integrations/stagehand.md): Trace Stagehand runs and watch the browser session recording synced to each act, extract, and LLM call in Tracevell.
- [Skyvern observability](https://tracevell.com/docs/integrations/skyvern.md): Trace Skyvern workflows and watch the browser session recording synced to each step, prompt, and screenshot in Tracevell.
- [Kernel observability](https://tracevell.com/docs/integrations/kernel.md): Trace computer-using agents that drive Kernel browsers, with the browser session recording synced to the agent's steps.
- [Playwright observability](https://tracevell.com/docs/integrations/playwright.md): Record Playwright browser sessions in JavaScript and Python and read them synced to the surrounding agent trace in Tracevell.
- [Puppeteer observability](https://tracevell.com/docs/integrations/puppeteer.md): Record Puppeteer browser sessions in JavaScript and read them synced to the surrounding agent trace in Tracevell.
- [Harbor agent evaluations](https://tracevell.com/docs/integrations/harbor.md): Add one flag to harbor run and read every trial as a scored datapoint with a full agent transcript, for Claude Code, Codex, or your own agent.
- [Langfuse observability](https://tracevell.com/docs/integrations/langfuse.md): Keep your Langfuse instrumentation and dual-export the same spans to Tracevell, then compare the two side by side before you switch.
- [OpenTelemetry](https://tracevell.com/docs/tracing/otel.md)
- [Understand Your Agent Using Tracevell Signals](https://tracevell.com/docs/guides/signals.md)
- [Tracing LLM calls in FastAPI with Tracevell](https://tracevell.com/docs/guides/fastapi.md)
- [Debug an AI Agent End to End: Record, Replay, Evaluate](https://tracevell.com/docs/guides/debugger-end-to-end.md)
- [Debug a Browser Use Agent with Tracevell](https://tracevell.com/docs/guides/browser-use-debugger.md)
- [Tracing AI SDK in Next.js with Tracevell](https://tracevell.com/docs/guides/nextjs-aisdk.md)
- [Tracing LLM calls in Next.js with Tracevell](https://tracevell.com/docs/guides/nextjs.md)
- [Evaluating LLM Tool Calls with Tracevell](https://tracevell.com/docs/guides/evaluating-tool-calls.md)
- [Self-hosting Tracevell](https://tracevell.com/docs/self-hosting/overview.md)
- [Self-host Tracevell with Docker Compose](https://tracevell.com/docs/self-hosting/docker-compose.md)
- [Self-host Tracevell on Kubernetes with Helm](https://tracevell.com/docs/self-hosting/kubernetes.md)
- [Helm Configuration Reference for Self-hosted Tracevell](https://tracevell.com/docs/self-hosting/configuration.md)
- [Observe](https://tracevell.com/docs/sdk/observe.md)
- [Manual Span Creation](https://tracevell.com/docs/sdk/manual-spans.md)
- [Span Methods](https://tracevell.com/docs/sdk/span-methods.md)
- [Trace Methods](https://tracevell.com/docs/sdk/trace-methods.md)
- [Context Utilities](https://tracevell.com/docs/sdk/context-utilities.md)
- [SDK Lifecycle](https://tracevell.com/docs/sdk/lifecycle.md)
- [Span Batching and Export Payload Size Limits](https://tracevell.com/docs/sdk/batching.md)
- [SDK Constants](https://tracevell.com/docs/sdk/constants.md)
- [Client](https://tracevell.com/docs/sdk/client.md)
- [Evaluations](https://tracevell.com/docs/sdk/evaluations.md)
- [TypeScript Instrumentation](https://tracevell.com/docs/sdk/typescript/instrumentation.md)
- [Python Instrumentation](https://tracevell.com/docs/sdk/python/instrumentation.md)
- [TypeScript Types](https://tracevell.com/docs/sdk/typescript/types.md)
- [Python Types](https://tracevell.com/docs/sdk/python/types.md)
- [Ingest OpenTelemetry traces](https://tracevell.com/docs/api-reference/ingestion/ingest-opentelemetry-traces.md): Accepts an OTLP ExportTraceServiceRequest encoded as protobuf or OTLP JSON.
- [Create spans](https://tracevell.com/docs/api-reference/ingestion/create-spans.md): Ingests a batch of spans using Tracevell's JSON span format.
- [Ingest OpenTelemetry logs](https://tracevell.com/docs/api-reference/ingestion/ingest-opentelemetry-logs.md): Accepts an OTLP ExportLogsServiceRequest encoded as protobuf. JSON-encoded OTLP is not accepted. This endpoint is in beta: ingested logs are stored in the `logs` table and can only be queried with SQL. They are not shown anywhere in the Tracevell UI, and Signals do not read them when analyzing traces.
- [Add labeling queue items](https://tracevell.com/docs/api-reference/labeling-queues/add-labeling-queue-items.md): Adds items to a labeling queue, optionally deduplicating them with idempotency keys.
- [Query project data](https://tracevell.com/docs/api-reference/sql/query-project-data.md): Executes a validated, read-only SQL query scoped to the authenticated project.
- [Get the current project](https://tracevell.com/docs/api-reference/projects/get-the-current-project.md): Returns the project ID associated with the supplied API key.
- [List datasets](https://tracevell.com/docs/api-reference/datasets/list-datasets.md): Lists project datasets, optionally filtered by dataset ID or name.
- [Create a dataset](https://tracevell.com/docs/api-reference/datasets/create-a-dataset.md): Creates an empty dataset. Dataset names do not need to be unique.
- [Get a dataset](https://tracevell.com/docs/api-reference/datasets/get-a-dataset.md): Returns a dataset by its canonical ID.
- [Delete a dataset](https://tracevell.com/docs/api-reference/datasets/delete-a-dataset.md): Deletes a dataset and its datapoints.
- [Update a dataset](https://tracevell.com/docs/api-reference/datasets/update-a-dataset.md): Renames a dataset. Dataset names do not need to be unique.
- [List dataset datapoints](https://tracevell.com/docs/api-reference/datasets/list-dataset-datapoints.md): Returns a paginated page of datapoints from a dataset identified by ID or name.
- [Create dataset datapoints](https://tracevell.com/docs/api-reference/datasets/create-dataset-datapoints.md): Adds datapoints to a dataset and can optionally create the dataset when identified by name.
- [List Signals](https://tracevell.com/docs/api-reference/signals/list-signals.md): Lists Signals in the project, optionally filtering by a name substring.
- [Create a Signal](https://tracevell.com/docs/api-reference/signals/create-a-signal.md): Creates a Signal with the first server-managed configuration version.
- [Get a Signal](https://tracevell.com/docs/api-reference/signals/get-a-signal.md): Returns a Signal in the authenticated project.
- [Delete a Signal](https://tracevell.com/docs/api-reference/signals/delete-a-signal.md): Deletes a Signal and returns its final state.
- [Update a Signal](https://tracevell.com/docs/api-reference/signals/update-a-signal.md): Updates a Signal. An update that changes the effective Signal configuration increments its server-managed version.
- [List LLM profiles](https://tracevell.com/docs/api-reference/llm-profiles/list-llm-profiles.md): Lists every LLM profile in the authenticated project's workspace.
- [Create an LLM profile](https://tracevell.com/docs/api-reference/llm-profiles/create-an-llm-profile.md): Creates an LLM profile in the authenticated project's workspace.
- [Get an LLM profile](https://tracevell.com/docs/api-reference/llm-profiles/get-an-llm-profile.md): Returns an LLM profile from the authenticated project's workspace. Secret values are masked.
- [Delete an LLM profile](https://tracevell.com/docs/api-reference/llm-profiles/delete-an-llm-profile.md): Deletes an LLM profile and returns its final state.
- [Update an LLM profile](https://tracevell.com/docs/api-reference/llm-profiles/update-an-llm-profile.md): Updates an LLM profile. Omitted fields retain their stored values, `models` replaces the complete model list, a supplied `config` replaces the whole stored config object, and supplied secrets are merged with stored secrets.
- [Create an evaluation](https://tracevell.com/docs/api-reference/evaluations/create-an-evaluation.md): Creates an evaluation run, generating a name when none is supplied.
- [Update an evaluation](https://tracevell.com/docs/api-reference/evaluations/update-an-evaluation.md): Updates an evaluation's name and/or metadata; omitted fields remain unchanged.
- [Add evaluation datapoints](https://tracevell.com/docs/api-reference/evaluations/add-evaluation-datapoints.md): Adds result datapoints to an evaluation run.
- [Update an evaluation datapoint](https://tracevell.com/docs/api-reference/evaluations/update-an-evaluation-datapoint.md): Updates a datapoint's executor output, scores, and optional trace association.
- [Changelog](https://tracevell.com/docs/changelog.md)

## OpenAPI Specs

- [openapi](/docs/openapi/openapi.yaml)


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