Guides
End-to-end walkthroughs that combine Confident AI features in implementation workflows.
Overview
Guides are end-to-end walkthroughs for common Confident AI implementation workflows.
Use these guides when you want to see how multiple Confident AI features work together in a real setup, with the relevant code and configuration in one place.
End-to-End Guides
Choose a guide based on the workflow you want to implement:
Build Test Runs from Traces
Open a test run, stream your app's traces into it as test cases, and let Confident AI evaluate each one at the trace and component level — over the API or OpenTelemetry.
Build Filtered Views from a URL
Construct a link that opens a Confident AI page with your filters already applied — share it, bookmark it, or generate it from CI.
Generating Reports for Stakeholders
Plan what leadership cares about, build it into a custom Executive Report template, and automate delivery so a curated report reaches your stakeholders on a schedule.
Evaluate MCP Servers
Set up distributed tracing for MCP hosts and servers, then evaluate whether your agent calls the right tools with the right arguments against a dataset before you ship.
Evaluating Your MCP Client in Code
Describe your MCP servers, record the tool calls your client made, and score them with deepeval from your own code so MCP tool calls show up apart from local functions.
Gating Red Teamed AI Agents with Governance Gates
Squeeze hundreds of agents through one standardized deployment gate, so no agent ships without current, correctly configured red teaming evidence.
Vibe Code Your Administration
Install the confident-client Agent Skill and drive the Confident AI SDK from Cursor, Claude Code, or Codex — creating projects, inviting members, and provisioning keys with plain-English prompts.
Provision Projects for Agents on the Fly
Use this guide to provision a dedicated Confident AI project for each agent your users build, then route its traces into the correct project.
Authenticate AI Connections with Client Credentials
Connect to an OAuth2-protected endpoint (Azure AD / Microsoft Entra ID or Auth0) by having Confident AI fetch a Bearer token before every request.
Set Up Long-Running AI Connections
Evaluate agents that take minutes to respond by enabling Async Responses on your AI connection and posting results back through the Confident API.
Set Up the Confident Agent in an Air-Gapped Environment
Deploy the Confident Agent in an air-gapped or egress-restricted network using outbound-only WebSocket Secure (WSS) connectivity.
Standardize Onboarding with Custom Agent Skills
Use this guide to serve project-specific onboarding and governance instructions to Claude Code, Codex, Cursor, and other coding agents.
Migration Guides
Move your traces from another platform with confident-trace. Every guide follows the same steps, with $200 in migration credits after a call:
Switch from Langfuse V1 to Confident AI
Move your traces from Langfuse SDK v1 to Confident AI with confident-trace. Auto instrumentation replaces your trace and generation objects.
Switch from Langfuse V2 to Confident AI
Move your traces from Langfuse SDK v2 to Confident AI with confident-trace. Your decorators, traces and trace fields map one to one.
Switch from Langfuse V3 to Confident AI
Move your traces from Langfuse SDK v3 to Confident AI with confident-trace. Your spans, provider clients and frameworks stay untouched.
Switch from Langfuse V4 to Confident AI
Move your OpenTelemetry traces from Langfuse SDK v4 to Confident AI with confident-trace. Your spans, provider clients and frameworks stay untouched.
Switch from LangSmith to Confident AI
Move your traces from LangSmith to Confident AI with confident-trace. Your LangChain and LangGraph spans, provider clients and frameworks stay untouched.
Switch from Arize to Confident AI
Move your traces from Arize Phoenix or Arize AX to Confident AI with confident-trace. Your OpenTelemetry spans and OpenInference instrumentors stay untouched.
Switch from Braintrust to Confident AI
Move your traces from Braintrust to Confident AI with confident-trace. Your OpenTelemetry spans, provider clients and frameworks stay untouched.
Last updated on