# Zespan

## Docs

- [AGENTS](https://docs.zespan.com/AGENTS.md)
- [API keys](https://docs.zespan.com/account/api-keys.md): Understand project API keys and personal API keys in Zespan, where to create and revoke them, and best practices for managing them safely.
- [Billing](https://docs.zespan.com/account/billing.md): Compare Zespan pricing plans, understand overage billing, view usage, and manage your subscription.
- [Organizations: workspaces, members, and settings](https://docs.zespan.com/account/organizations.md): Learn how to create and manage organizations in Zespan, invite team members, assign roles, and control access to your projects and billing.
- [The raw dependency graph for a project](https://docs.zespan.com/api-reference/blast-radius/the-raw-dependency-graph-for-a-project.md): Returns every node and edge in the project's dependency graph, optionally filtered to a subset of node kinds — the same graph `GET /v1/projects/{id}/blast-radius` traverses from a single root, but unfiltered by root. Sized for a future graph visualization; nothing in the dashboard renders it today.
- [What breaks if this resource changes](https://docs.zespan.com/api-reference/blast-radius/what-breaks-if-this-resource-changes.md): Walks the dependency graph from one root node and returns every dependent found (breadth-first, bounded by `maxDepth`/`maxNodes`), summarized into an impact rollup. Backs the pre-release impact card on the Prompts page and the delete-blocking check on Evaluations.
- [Coverage preview for a framework and period, without generating](https://docs.zespan.com/api-reference/compliance/coverage-preview-for-a-framework-and-period-without-generating.md): Runs the same evidence queries a control-evidence document would for the given framework and period, and reports per-control record counts — without rendering or storing anything. This is what the Compliance page calls before you commit to generating a pack, so a gap in the underlying data is visibl…
- [Generate an evidence pack](https://docs.zespan.com/api-reference/compliance/generate-an-evidence-pack.md): Creates an `EvidencePack` row in `pending` status and enqueues generation on a background worker — this endpoint returns immediately, before the document exists. Poll `GET /v1/projects/{id}/evidence-packs/{packId}` (or list packs) to watch `status` move through `processing` to `completed` (or `faile…
- [Get one evidence pack, optionally its content](https://docs.zespan.com/api-reference/compliance/get-one-evidence-pack-optionally-its-content.md): Returns the pack row plus, once `status` is `completed`, either a presigned `downloadUrl` (object-storage backend) or the raw `content` string (local-disk backend, when `download=true` is passed) — never both. A pack belonging to a different project than `id` returns `404`, the same non-enumerable b…
- [List available compliance frameworks](https://docs.zespan.com/api-reference/compliance/list-available-compliance-frameworks.md): A static capability listing — the frameworks Zespan can map evidence to today, each with its controls. `soc2` is the only registered framework as of this release; EU AI Act and ISO/IEC 42001 are not yet available. Not project-scoped and not plan-gated: a customer deciding whether to upgrade needs to…
- [List evidence packs for a project](https://docs.zespan.com/api-reference/compliance/list-evidence-packs-for-a-project.md): Paginated list, newest first. `evidenceIndex` (the citation list) is deliberately excluded from every row here — it can hold hundreds of entries per pack and this view never renders them; fetch a single pack to read it.
- [Re-verify a generated evidence pack](https://docs.zespan.com/api-reference/compliance/re-verify-a-generated-evidence-pack.md): Re-hashes the stored document and re-resolves every citation it made against live data, scoped to the pack's project or organization. Never errors on a data condition — a deleted record, a deleted storage object, or a pack that never finished generating are all reported as *results*, not exceptions.…
- [Compare two dataset runs](https://docs.zespan.com/api-reference/datasets/compare-two-dataset-runs.md): Compare two runs of a dataset. Returns each dataset item shared by both runs with each run's trace id and score side by side, plus each run's average score over the shared items.
- [Create or fetch a dataset run](https://docs.zespan.com/api-reference/datasets/create-or-fetch-a-dataset-run.md): Create a named run for a dataset, or return the existing run with that name. Idempotent: the SDK calls this every time a job starts.
- [Get dataset run detail](https://docs.zespan.com/api-reference/datasets/get-dataset-run-detail.md): Fetch a run with every linked item joined to its dataset item content and, if the run has been scored, its evaluation result.
- [Link an item to a dataset run](https://docs.zespan.com/api-reference/datasets/link-an-item-to-a-dataset-run.md): Link a dataset item to a run by recording the trace produced for it. Idempotent: re-linking updates the stored trace pointer.
- [List dataset items](https://docs.zespan.com/api-reference/datasets/list-dataset-items.md): List the items in a dataset (up to 1000), oldest first.
- [List dataset runs](https://docs.zespan.com/api-reference/datasets/list-dataset-runs.md): List runs for a dataset (newest first), each enriched with its latest scoring status and average score if it has been scored.
- [Resolve a dataset by name](https://docs.zespan.com/api-reference/datasets/resolve-a-dataset-by-name.md): Resolve a dataset id from its name within the authenticated project. This is the SDK's entry point, since customer code refers to datasets by name.
- [Run a dataset against a registered HTTP Target](https://docs.zespan.com/api-reference/datasets/run-a-dataset-against-a-registered-http-target.md): Start a Zespan-executed run: for every item in the dataset, Zespan hydrates the target's request template with the item's `input` and POSTs it directly to the registered HTTP Target's endpoint, capturing the raw response as a trace tagged `sdk_name: "zespan-http-endpoint"`. Unlike every other datase…
- [Score a dataset run](https://docs.zespan.com/api-reference/datasets/score-a-dataset-run.md): Trigger scoring of a run's linked traces with an evaluator. Creates a fresh evaluation run over the run's trace ids and enqueues it for asynchronous scoring.
- [Run a guardrails check](https://docs.zespan.com/api-reference/guardrails/run-a-guardrails-check.md): Evaluate text against the authenticated project's enabled guardrails at runtime. Returns whether the text is allowed, the per-guardrail results, and any modified (e.g. redacted) text. Requires `x-api-key`.
- [Ingest trace events](https://docs.zespan.com/api-reference/ingestion/ingest-trace-events.md): Ingest a batch of trace/span events. The request body is newline-delimited JSON (`application/x-ndjson`): one JSON event object per line. A maximum of 100 events per request and a 1 MB body size limit apply. Events are validated and queued asynchronously; invalid lines are skipped rather than failin…
- [API reference](https://docs.zespan.com/api-reference/introduction.md): Authenticate, choose a base URL, and call the Zespan Public API directly or through the SDKs.
- [Dismiss a model deprecation finding](https://docs.zespan.com/api-reference/model-lifecycle/dismiss-a-model-deprecation-finding.md): Suppresses a finding at its current urgency band. It reopens automatically — exactly once, and notifies again — the next time the daily scan finds the band has tightened (90 → 30 → 7 days, or into retired). A band that widens, which only happens when a feed correction pushes the retirement date furt…
- [List model deprecation findings for a project](https://docs.zespan.com/api-reference/model-lifecycle/list-model-deprecation-findings-for-a-project.md): Returns findings from the daily deprecation scan for this project, soonest deadline first (an already-retired model, with a negative `daysRemaining`, sorts to the very top). Backs the Model Lifecycle dashboard page, the Overview widget, and the Models page banner.
- [The full bundled model lifecycle feed](https://docs.zespan.com/api-reference/model-lifecycle/the-full-bundled-model-lifecycle-feed.md): Returns Zespan's curated, bundled catalogue of provider-announced deprecation and retirement dates in full — not scoped to a project, since the catalogue is the same for every tenant. This is the raw data [Model Lifecycle](/dashboard/model-lifecycle) findings and the Models page's Lifecycle column a…
- [Per-model usage, cost, latency, and error rate](https://docs.zespan.com/api-reference/models/per-model-usage-cost-latency-and-error-rate.md): Aggregates every model called in the project over the selected range, ranked by the chosen sort. Backs the Models dashboard page's table.
- [Export OTLP logs (not implemented)](https://docs.zespan.com/api-reference/opentelemetry/export-otlp-logs-not-implemented.md): Not implemented. This endpoint authenticates the API key and returns `501`; no logs are stored. Send traces to `/v1/traces` instead.
- [Export OTLP metrics (not implemented)](https://docs.zespan.com/api-reference/opentelemetry/export-otlp-metrics-not-implemented.md): Not implemented. This endpoint authenticates the API key and returns `501`; no metrics are stored. Send traces to `/v1/traces` instead.
- [Export OTLP traces](https://docs.zespan.com/api-reference/opentelemetry/export-otlp-traces.md): OTLP-compatible trace export endpoint. Accepts an `ExportTraceServiceRequest` payload as JSON (`application/json`) or protobuf (`application/x-protobuf`). Spans are converted to Zespan events and queued asynchronously.
- [Distinct outcome kinds reported for a project](https://docs.zespan.com/api-reference/outcomes/distinct-outcome-kinds-reported-for-a-project.md): Returns the distinct `kind` values reported for this project in `[now - range, now]`, sorted alphabetically. Used to populate kind filters in the dashboard.
- [Outcome summary by agent or model, joined to cost](https://docs.zespan.com/api-reference/outcomes/outcome-summary-by-agent-or-model-joined-to-cost.md): Returns one row per distinct value of `dimension` (agent name or model name), summarizing every outcome reported in `[now - range, now]`: total outcomes, successes, total value, and the real LLM cost of the traces those outcomes are attributed to, plus two derived ratios (`costPerSuccess`, `valuePer…
- [Report business outcomes attributed to traces](https://docs.zespan.com/api-reference/outcomes/report-business-outcomes-attributed-to-traces.md): Report a batch of business outcomes (a deflected ticket, an avoided refund, an SLA met) attributed to traces your SDK already emitted. This is a direct, synchronous write to ClickHouse — not routed through the same async queue as `POST /v1/ingest` — since it's a single low-volume out-of-band call ra…
- [Create a prompt version](https://docs.zespan.com/api-reference/prompts/create-a-prompt-version.md): Create a new version of a prompt. If the named prompt already exists a new incremented version is created; otherwise version 1 is created. The `latest` label is always applied automatically.
- [Get a prompt](https://docs.zespan.com/api-reference/prompts/get-a-prompt.md): Fetch a prompt by name. Without `version` or `label` the latest version is returned. The response includes `resolvedPrompt` with prompt dependencies inlined.
- [List prompt folders](https://docs.zespan.com/api-reference/prompts/list-prompt-folders.md): List the distinct folder paths used across the project's prompts.
- [List prompt versions](https://docs.zespan.com/api-reference/prompts/list-prompt-versions.md): List every version of a prompt family, newest first.
- [List prompts](https://docs.zespan.com/api-reference/prompts/list-prompts.md): List prompts for the authenticated project. When using an API key the project is inferred from the key, so `projectId` is optional.
- [Move a prompt to a folder](https://docs.zespan.com/api-reference/prompts/move-a-prompt-to-a-folder.md): Move a prompt family into a folder path, or to the root by passing `null`. The folder is applied to every version of the named prompt.
- [Set labels on a prompt version](https://docs.zespan.com/api-reference/prompts/set-labels-on-a-prompt-version.md): Replace the set of labels on a specific prompt version. Assigning the `production` label promotes that version and triggers a background quality regression check.
- [Set tags on a prompt family](https://docs.zespan.com/api-reference/prompts/set-tags-on-a-prompt-family.md): Replace the set of tags across all versions of a prompt family.
- [Check whether spans are actually arriving for a project](https://docs.zespan.com/api-reference/sdk-cli-support/check-whether-spans-are-actually-arriving-for-a-project.md): Reports the most recent span timestamp, a 24h span count, and any recorded ingest rejections for a project — the data `zespan doctor`'s Data flow check reads to tell "your integration is broken" apart from "this project hasn't sent its first trace yet."
- [Resolve project identity from an API key](https://docs.zespan.com/api-reference/sdk-cli-support/resolve-project-identity-from-an-api-key.md): Answers "which project does this API key belong to?" — the one thing `GET /v1/sdk/config` can't answer, since that route requires the caller to already know the `projectId` and rejects a mismatch. This is `zespan doctor`'s first call, and how it establishes project identity before anything else in t…
- [zespan doctor](https://docs.zespan.com/cli/doctor.md): Diagnose SDK setup issues — bad API key, unreachable API, no data arriving, or a provider client wrapped in the wrong order — before they turn into a churned trial.
- [CLI overview](https://docs.zespan.com/cli/overview.md): Install @zespan/cli, understand its two binaries (zespan and zespan-gate), and configure it with flags, environment variables, or a committed .zespan.yaml.
- [Evidence Packs — audit-ready compliance documents](https://docs.zespan.com/compliance/evidence-packs.md): Generate hash-addressed, re-verifiable audit documents — a per-agent Compliance Card or a SOC 2 control evidence report — from your existing Zespan data.
- [Frameworks and controls](https://docs.zespan.com/compliance/frameworks.md): The SOC 2 control mapping evidence packs draw on today — CC6.1, CC7.2, CC8.1 — what each control's evidence comes from, and what's not yet available.
- [Verification](https://docs.zespan.com/compliance/verification.md): What GET /v1/projects/:id/evidence-packs/:packId/verify re-checks against a generated evidence pack — and, importantly, what it can't check at all.
- [Core concepts: the Zespan data model](https://docs.zespan.com/concepts.md): Learn how Zespan structures agent observability data — events, spans, span kinds, traces, projects, and organizations — and how cost is calculated from token counts.
- [Agent Registry](https://docs.zespan.com/dashboard/agent-registry.md): A live map of every agent in your system, the tools they use, and how they connect to each other.
- [Agents](https://docs.zespan.com/dashboard/agents.md): Per-agent performance dashboards — cost, latency, error rate, and tool usage broken down by agent.
- [Anomaly Detection — catch metric deviations before they become incidents](https://docs.zespan.com/dashboard/ai-features.md): Run on-demand statistical scans over your LLM traffic to detect latency spikes, error rate surges, and cost drifts — with plain-English explanations and automatic incident correlation.
- [Alerts](https://docs.zespan.com/dashboard/alerts.md): Create threshold rules on error rate, cost, latency, and eval scores. Get notified via email, Slack, PagerDuty, OpsGenie, Discord, Jira, Freshservice, or webhook.
- [Annotation Queues](https://docs.zespan.com/dashboard/annotation-queues.md): Route a filtered set of traces to a human reviewer for manual pass/fail or score annotation, stored alongside automated judge scores.
- [Blast Radius — what breaks if you change this](https://docs.zespan.com/dashboard/blast-radius.md): A dependency graph across prompts, agents, models, guardrails, evaluators, and alerts, backing an impact check shown before you release a prompt and a delete-blocking check on evaluators.
- [Changes — a project-wide timeline of what changed and when](https://docs.zespan.com/dashboard/changes.md): A single feed of every prompt deploy, agent promotion, policy edit, and pipeline deploy in your project, so 'what changed?' is a page you visit instead of a question you Slack someone.
- [Costs — understand and control your LLM spend](https://docs.zespan.com/dashboard/costs.md): Break down your LLM spend by model and over time, track cache savings, see your month-to-date total, and get a 30-day forecast to avoid billing surprises.
- [Datasets](https://docs.zespan.com/dashboard/datasets.md): Manage evaluation datasets — create from traces, upload CSV, and score your own pipeline's runs against them.
- [Environments — filter traces, evaluations, and guardrails by deployment stage](https://docs.zespan.com/dashboard/environments.md): Every project ships with dev, staging, and prod. Add your own, switch between them from the header, and filter any view down to just one.
- [Errors](https://docs.zespan.com/dashboard/errors.md): Track, group, and triage errors across your agents — model errors, tool failures, timeouts, and rate limits.
- [Evaluations — measure and trend custom LLM metrics](https://docs.zespan.com/dashboard/evaluations.md): Define custom evaluation metrics, run them against your traces, and track results over time to detect quality regressions and measure the impact of prompt changes.
- [Guardrails — configure content safety policies](https://docs.zespan.com/dashboard/guardrails.md): Create and manage guardrail policies in the Zespan dashboard to block, redact, or flag unsafe LLM inputs and outputs without redeploying your application.
- [HTTP Targets](https://docs.zespan.com/dashboard/http-targets.md): Register an externally-hosted agent endpoint so a dataset run can call it directly, without instrumenting it with the Zespan SDK.
- [Incidents — detect, correlate, and resolve LLM issues](https://docs.zespan.com/dashboard/incidents.md): The Incidents page surfaces active and resolved issues detected across your LLM traffic, correlates related anomalies and errors, and tracks resolution state.
- [Issues — recurring failures grouped automatically](https://docs.zespan.com/dashboard/issues.md): Zespan clusters repeated failed and degraded traces into a single recurring Issue, using the same deterministic verdict system shown on every trace, so you stop re-discovering the same problem one trace at a time.
- [Model Lifecycle — provider deprecation radar](https://docs.zespan.com/dashboard/model-lifecycle.md): Detects when a model you're actually calling has a provider-announced end-of-life, with real measured call volume, cost, and affected agents — never a fabricated quality or regression comparison.
- [Models — per-model usage, reliability, and latency](https://docs.zespan.com/dashboard/models.md): A per-model table of usage, cost, latency, error rate, and agent adoption, plus a Lifecycle column flagging models with a provider-announced end-of-life.
- [System Health — your agent health at a glance](https://docs.zespan.com/dashboard/overview.md): The System Health page shows a verdict banner judged against your own thresholds, golden-signal KPI cards, a triage row, and five drill-down tabs — all in one place.
- [Performance — latency metrics and throughput analysis](https://docs.zespan.com/dashboard/performance.md): Track P50, P90, and P99 latency, time-to-first-token for streaming calls, and request throughput across models, environments, and time ranges.
- [Playground — compare a prompt across models side-by-side](https://docs.zespan.com/dashboard/playground.md): Run one versioned prompt against up to three models at once, compare cost, latency, and output, and save the winner as a new prompt version.
- [Prompts — version, deploy, and gate your prompt library](https://docs.zespan.com/dashboard/prompts.md): Store, version, label, and deploy prompt templates through a quality gate. Fetch prompts at runtime via the SDK so you can update them without redeploying.
- [Sessions — track multi-turn user conversations](https://docs.zespan.com/dashboard/sessions.md): Use session IDs to group multiple LLM calls from a single user conversation and analyze session-level cost, latency, and turn count in the Sessions dashboard.
- [Simulations — test LLM behavior before shipping](https://docs.zespan.com/dashboard/simulations.md): Run your LLM application against a dataset of test cases to catch regressions before they reach production. Compare outputs, scores, and latency across prompt versions.
- [Tools](https://docs.zespan.com/dashboard/tools.md): Track every tool your agents call — usage counts, failure rates, latency, and argument/return value inspection.
- [Traces — inspect every agent run and its steps](https://docs.zespan.com/dashboard/traces.md): Browse your full agent run log, filter by model or status, and drill into any run: a plain-English health verdict, three evaluation lenses, a span navigator + inspector, and flow and flame graphs of every agent turn, tool call, and LLM interaction.
- [Value — business outcomes attributed to agents and models](https://docs.zespan.com/dashboard/value.md): See which agents and models actually drive business results — deflected tickets, avoided refunds, SLAs met — reported against your traces and summarized by success rate, value, cost, and value per dollar.
- [ZespanPilot — AI copilot for your LLM operations](https://docs.zespan.com/dashboard/zespanpilot.md): ZespanPilot is an action-first AI copilot that manages alerts, guardrails, prompts, and SDK configuration through natural language conversation, with full audit logging.
- [Environments — separate production, staging, and development data](https://docs.zespan.com/guides/environments.md): Use the environment tag and project structure to cleanly separate LLM traces from different deployment environments in Zespan.
- [Production checklist](https://docs.zespan.com/guides/production-checklist.md): Everything to verify before shipping Zespan to production: API key security, sampling, flush handling, alert rules, and data privacy settings.
- [Serverless deployment guide](https://docs.zespan.com/guides/serverless.md): Ensure Zespan events are reliably sent from AWS Lambda, Vercel Functions, Netlify Functions, and Google Cloud Run by handling flush correctly in each environment.
- [Testing and mocking](https://docs.zespan.com/guides/testing-and-mocking.md): Disable Zespan in tests, mock the SDK for unit testing, and configure CI environments.
- [Webhooks — receive real-time alert notifications](https://docs.zespan.com/guides/webhooks.md): Configure Zespan to POST a signed JSON payload to your endpoint when an alert fires. Covers payload schema, HMAC signature verification, delivery guarantees, and integration examples.
- [Zespan MCP server](https://docs.zespan.com/guides/zespan-mcp.md): Connect Claude Desktop, Cursor, or any MCP client to Zespan to query your traces, metrics, and errors in natural language.
- [Zespan: Agent Observability for Production AI](https://docs.zespan.com/index.md): Zespan gives you complete visibility into every agent run, tool call, and LLM interaction. Trace multi-agent systems, monitor costs, debug failures, and optimize your AI operations in one dashboard.
- [Zespan: AI Agent Reliability Platform](https://docs.zespan.com/introduction.md): Zespan doesn't just watch your AI agents — it blocks unsafe runs mid-execution from inside your own process, classifies every trace pass/fail with a deterministic verdict, and clusters recurring failures automatically. Engineering teams ship agents they can trust, not just agents they can watch fail…
- [Audit log](https://docs.zespan.com/platform/audit.md): A tamper-evident log of every action taken in your Zespan organization — who did what, and when.
- [Environment variables reference](https://docs.zespan.com/platform/environment-variables.md): Complete reference for all environment variables used by the Zespan API server, including database connections, auth configuration, AI providers, and integrations.
- [LLM Connections (BYOK)](https://docs.zespan.com/platform/llm-connections.md): Connect your own LLM provider keys per project to run Playground, Evaluations, and Simulations on your own inference spend.
- [Security](https://docs.zespan.com/platform/security.md): How Zespan protects your data: encryption, authentication, tenant isolation, PII controls, audit logging, and data deletion.
- [Self-hosting Zespan](https://docs.zespan.com/platform/self-hosting.md): Deploy Zespan on your own infrastructure with an Enterprise license. Full data sovereignty — all traces, metrics, and AI analysis stay within your network.
- [Single sign-on (SSO)](https://docs.zespan.com/platform/sso.md): Connect an OIDC identity provider so your team signs in through your existing SSO, verify domain ownership, and optionally enforce SSO-only login.
- [Quickstart: instrument your first agent in 5 minutes](https://docs.zespan.com/quickstart.md): Send your first agent trace to Zespan by wrapping your LLM client or agent framework with the SDK. No manual instrumentation required — two lines of setup.
- [Changelog](https://docs.zespan.com/reference/changelog.md): Release history for the Zespan platform and SDKs.
- [Compliance evidence limitations](https://docs.zespan.com/reference/compliance-limitations.md): The permanent data gaps evidence packs report about themselves, quoted from the exact limitations text a generated document prints.
- [Model lifecycle feed](https://docs.zespan.com/reference/model-lifecycle-feed.md): The curated, bundled catalogue of provider-announced model deprecation and retirement dates backing Model Lifecycle — its field reference, how it's maintained, and how to report a wrong or missing entry.
- [Supported models and pricing](https://docs.zespan.com/reference/models.md): All LLM models supported by Zespan's built-in pricing table, with per-token costs for input, output, and cached tokens used in cost_usd calculation.
- [Span kinds reference](https://docs.zespan.com/reference/span-kinds.md): Complete reference for all Zespan span kinds — what each represents, when it is emitted, and how it appears in the trace flame graph and agent registry.
- [Agent tracing — withAgent and multi-agent systems](https://docs.zespan.com/sdk/agent-tracing.md): Use withAgent (TypeScript) or with_agent (Python) to trace multi-agent workflows, capture planning steps, tool calls, and agent-to-agent handoffs as linked spans in the Zespan dashboard.
- [CI quality gate — zespan-gate CLI](https://docs.zespan.com/sdk/cli.md): Gate a merge or deploy on a prompt version's evaluation quality using @zespan/cli's zespan-gate binary — a dependency-free wrapper around the same quality-gate route the dashboard's Versions tab uses.
- [SDK config propagation — update live applications without redeployment](https://docs.zespan.com/sdk/config-propagation.md): Understand how ZespanPilot pushes SDK configuration changes to running applications in real time via the ingest response config version field.
- [Dataset runs — running your pipeline over a dataset with the SDK](https://docs.zespan.com/sdk/dataset-runs.md): Fetch a dataset's items, run your own prompt or agent against each one, and link the resulting traces back as a named run using DatasetsClient — so an evaluator can score it and gate quality. TypeScript and Python.
- [FastAPI middleware — automatic HTTP request tracing](https://docs.zespan.com/sdk/fastapi-middleware.md): Add ZespanFastAPIMiddleware to your FastAPI app to automatically trace every HTTP request and instrument individual LLM calls with @observe_llm and @observe_span.
- [Flask middleware — automatic HTTP request tracing](https://docs.zespan.com/sdk/flask-middleware.md): Add ZespanFlaskExtension to your Flask app to automatically trace every HTTP request and instrument LLM calls with @observe_llm and @observe_span decorators.
- [Guardrails — content checking for LLM calls](https://docs.zespan.com/sdk/guardrails.md): Enable pre- and post-LLM content checks on any wrapped client to block, redact, or flag unsafe inputs and outputs using your configured guardrail policies.
- [Agno](https://docs.zespan.com/sdk/integrations/agno.md): Trace Agno multi-agent teams via OpenTelemetry auto-instrumentation, no Zespan SDK required.
- [Anthropic](https://docs.zespan.com/sdk/integrations/anthropic.md): Trace Anthropic Claude API calls automatically in TypeScript and Python — messages, tool use, and streaming — with wrapAnthropic() or patch_anthropic().
- [AutoGen / AG2](https://docs.zespan.com/sdk/integrations/autogen.md): Trace AutoGen agent conversations with Zespan. Wrap individual agents to capture replies, latency, and guardrail checks.
- [AWS Bedrock](https://docs.zespan.com/sdk/integrations/bedrock.md): Trace Amazon Bedrock model invocations in TypeScript and Python with wrapBedrock() or patch_bedrock().
- [Chroma](https://docs.zespan.com/sdk/integrations/chroma.md): Auto-trace Chroma collection query, add, and upsert calls with zespan.autopatch() — retrieval spans and RAG context capture with no code changes to your Chroma client calls.
- [Cloudflare AI Gateway](https://docs.zespan.com/sdk/integrations/cloudflare-ai-gateway.md): Trace every LLM call passing through Cloudflare AI Gateway via its built-in OpenTelemetry exporter, no Zespan SDK required.
- [Cohere](https://docs.zespan.com/sdk/integrations/cohere.md): Trace Cohere chat API calls in TypeScript and Python with wrapCohere() or patch_cohere().
- [CrewAI](https://docs.zespan.com/sdk/integrations/crewai.md): Trace CrewAI crew executions with Zespan. Wrap your Crew instance to capture agent runs, task outputs, latency, and guardrail checks.
- [Custom integrations](https://docs.zespan.com/sdk/integrations/custom.md): Don't see your framework in the list? Instrument it yourself with startSpan/start_span and get full tracing, cost tracking, and evaluation scores today — no official integration required.
- [DSPy](https://docs.zespan.com/sdk/integrations/dspy.md): Trace DSPy modules and optimizers via OpenTelemetry auto-instrumentation, no Zespan SDK required.
- [Google ADK](https://docs.zespan.com/sdk/integrations/google-adk.md): Trace Google Agent Development Kit (ADK) agents and multi-agent systems using instrumentADK, wrapADKRunner, and wrapADKAgent in TypeScript and Python.
- [Google Generative AI](https://docs.zespan.com/sdk/integrations/google-genai.md): Trace Gemini text, image generation, embeddings, and Veo video generation in TypeScript and Python — works with both the legacy and new Google Gemini SDKs.
- [Groq](https://docs.zespan.com/sdk/integrations/groq.md): Trace Groq inference calls in TypeScript and Python — optimized for high-throughput, low-latency workloads — with wrapGroq() or patch_groq().
- [Haystack](https://docs.zespan.com/sdk/integrations/haystack.md): Trace Haystack pipeline runs with Zespan using the built-in tracer integration.
- [Kong AI Gateway](https://docs.zespan.com/sdk/integrations/kong-ai-gateway.md): Trace every LLM call passing through Kong AI Gateway via its OpenTelemetry plugin, no Zespan SDK required.
- [LangChain](https://docs.zespan.com/sdk/integrations/langchain.md): Trace LangChain chains, agents, tools, and retrievers automatically using ZespanCallbackHandler — covering TypeScript and Python with full span-level detail.
- [LiteLLM](https://docs.zespan.com/sdk/integrations/litellm.md): Trace LiteLLM calls across any provider in TypeScript and Python with wrapLiteLLM() or patch_litellm().
- [LiteLLM Proxy (Gateway)](https://docs.zespan.com/sdk/integrations/litellm-proxy.md): Trace every call through a self-hosted LiteLLM Proxy gateway via its built-in OpenTelemetry callback, no Zespan SDK required.
- [LlamaIndex](https://docs.zespan.com/sdk/integrations/llamaindex.md): Trace LlamaIndex LLM calls, tool use, and agent steps with Zespan using the callback handler.
- [Mastra](https://docs.zespan.com/sdk/integrations/mastra.md): Trace Mastra agents and workflows via its built-in OpenTelemetry exporter, no Zespan SDK required.
- [Mistral](https://docs.zespan.com/sdk/integrations/mistral.md): Trace Mistral AI API calls in TypeScript and Python with wrapMistral() or patch_mistral().
- [Ollama](https://docs.zespan.com/sdk/integrations/ollama.md): Trace local Ollama models via OpenTelemetry auto-instrumentation, no Zespan SDK required.
- [OpenAI](https://docs.zespan.com/sdk/integrations/openai.md): Trace OpenAI chat completions automatically in TypeScript and Python — streaming, tool calls, guardrails, and config-driven retries and fallbacks, with two lines of code.
- [OpenAI Agents SDK](https://docs.zespan.com/sdk/integrations/openai-agents-sdk.md): Trace OpenAI's Agents SDK (agent runs, handoffs, tool calls) via OpenTelemetry auto-instrumentation, no Zespan SDK required.
- [OpenRouter](https://docs.zespan.com/sdk/integrations/openrouter.md): Trace calls made through OpenRouter's unified API in TypeScript and Python with wrapOpenRouter() or patch_openrouter().
- [Integrations overview](https://docs.zespan.com/sdk/integrations/overview.md): What an 'integration' means in Zespan — auto-instrumentation for LLM providers, agent frameworks, RAG frameworks, and vector databases — and where to start for your stack.
- [Pinecone](https://docs.zespan.com/sdk/integrations/pinecone.md): Auto-trace Pinecone query and upsert calls with zespan.autopatch() — retrieval spans and RAG context capture with no code changes to your Pinecone client calls.
- [Pydantic AI](https://docs.zespan.com/sdk/integrations/pydantic-ai.md): Configure a Python Pydantic AI agent to export its OpenTelemetry/Logfire traces to Zespan, using an environment variable helper from the Zespan TypeScript SDK.
- [Qdrant](https://docs.zespan.com/sdk/integrations/qdrant.md): Auto-trace Qdrant query_points/search and upsert calls with zespan.autopatch() — retrieval spans and RAG context capture with no code changes to your Qdrant client calls.
- [Semantic Kernel](https://docs.zespan.com/sdk/integrations/semantic-kernel.md): Trace Microsoft Semantic Kernel LLM calls and plugin executions with Zespan.
- [Strands Agents](https://docs.zespan.com/sdk/integrations/strands-agents.md): Trace AWS Strands Agents via its built-in OpenTelemetry support, no Zespan SDK required.
- [Vectra](https://docs.zespan.com/sdk/integrations/vectra.md): Trace vectra-js / vectra-py RAG pipelines with ZespanVectraCallbackHandler — ingestion, retrieval, reranking, and generation.
- [Vercel AI SDK](https://docs.zespan.com/sdk/integrations/vercel-ai.md): Trace Vercel AI SDK calls — generateText, streamText, generateObject, tool calls — with Zespan via OpenTelemetry.
- [Weaviate](https://docs.zespan.com/sdk/integrations/weaviate.md): Auto-trace Weaviate near_vector query and data insert calls with zespan.autopatch() — retrieval spans and RAG context capture with no code changes to your Weaviate v4/v3 client calls.
- [Manual spans — tracing custom operations](https://docs.zespan.com/sdk/manual-spans.md): Use startSpan/start_span to instrument RAG pipelines, custom model wrappers, evaluation harnesses, and any non-LLM operation you want to appear in your Zespan traces — in TypeScript or Python.
- [OpenTelemetry integration](https://docs.zespan.com/sdk/otel-integration.md): Send traces from any OTel-instrumented app straight to Zespan with no Zespan SDK at all, or use initOTel/init_otel, getTracer/get_tracer, createSpan/create_span, and withSpan/with_span to dual-export Zespan-instrumented spans to your own OTel backend.
- [Outcomes — reporting business results back to a trace](https://docs.zespan.com/sdk/outcomes.md): Use outcome() to attribute a business result — a deflected ticket, an avoided refund, an SLA met — back to the trace, agent, and model that produced it, usually minutes or hours after the trace ended.
- [SDK overview](https://docs.zespan.com/sdk/overview.md): Zespan ships a TypeScript SDK and a Python SDK with the same feature set. This page maps what each one does and where to go for a specific feature, instead of one giant per-language reference.
- [PII redaction — protect sensitive data in traces](https://docs.zespan.com/sdk/pii-redaction.md): Configure Zespan's built-in PII redaction — key-based and pattern-based — to remove sensitive values from captured LLM data before it's sent to the ingest endpoint.
- [Prompt management — versioned prompts with the SDK](https://docs.zespan.com/sdk/prompt-management.md): Fetch, compile, create, and manage versioned prompts from the Zespan prompt library using PromptClient, with built-in caching and variable substitution. TypeScript and Python.
- [Python SDK — zespan](https://docs.zespan.com/sdk/python.md): Install and configure the Zespan Python SDK to trace OpenAI, Anthropic, Google, Bedrock, Mistral, Groq, and LiteLLM calls, manage agent workflows, and flush events in serverless environments.
- [TypeScript SDK — @zespan/sdk](https://docs.zespan.com/sdk/typescript.md): Install and configure the Zespan TypeScript SDK to trace OpenAI, Anthropic, Google, Bedrock, Mistral, Groq, and LiteLLM calls, manage agent workflows, and flush events in serverless environments.
- [Troubleshooting](https://docs.zespan.com/troubleshooting.md): Solutions to the most common problems when integrating Zespan with your agents — from missing traces to cost calculation issues and broken span trees.

## OpenAPI Specs

- [openapi](https://docs.zespan.com/api-reference/openapi.yaml)