> For the complete documentation index, see [llms.txt](/llms.txt). > A full single-fetch corpus is available at [llms-full.txt](/llms-full.txt). --- title: AGNT5 learning path description: Follow a practical sequence from your first local workflow to deployment, operations, and improvement. last_verified: 2026-10-09 --- Use this path when you are new to AGNT5 and want a clear order through the docs. AGNT5 is organized around three jobs: build durable agentic workflows, run them in managed environments, and improve them from production evidence. ## The path at a glance | Stage | What you do | Start here | |---|---|---| | [1. Run a working workflow](#1-run-a-working-workflow) | Install the CLI and run your first workflow | [Quickstart](/docs/quickstart.md) | | [2. Learn the build model](#2-learn-the-build-model) | Pick the right primitive: workflow, function, agent, tool | [Workflows](/docs/build/workflows.md) | | [3. Build locally](#3-build-locally) | Iterate with hot reload, approvals, and events | [Local development](/docs/build/local-development.md) | | [4. Go live](#4-go-live) | Deploy a worker and promote it through environments | [Deploying](/docs/run/deploying.md) | | [5. Operate with trust](#5-operate-with-trust) | Debug production runs with traces, logs, and metrics | [Runs](/docs/run/deploying.md#runs) | | [6. Improve behavior](#6-improve-behavior) | Measure changes with datasets, scorers, and experiments | [Datasets](/docs/improve/datasets.md) | ## 1. Run a working workflow Get a project running and see its **trace**: the recorded steps, model calls, timing, and cost of one run. 1. [Install the CLI](/docs/install-cli.md): install `agnt5`, verify the binary, and authenticate. 2. [Hacker News digest quickstart](/docs/quickstart.md): create a workflow project, run it in an AGNT5 Cloud dev environment, and inspect the trace. Prefer to have a coding agent do the setup? Use [Build with AI](/docs/build-with-ai.md) instead of step 2. Outcome: you have seen a workflow run and can inspect what happened. ## 2. Learn the build model A **workflow** is a durable program, a **function** is a retryable unit of work, an **agent** reasons and calls tools, and a **prompt** keeps model instructions versioned with code. - [Workflows](/docs/build/workflows.md): checkpointed steps, durable sleep, parallel work, state, and triggers. - [Functions](/docs/build/functions.md): deterministic work with retries, backoff, timeouts, and trace visibility. - [Agents](/docs/build/agents.md): model instructions, tools, handoffs, and loop limits. - [Tools](/docs/build/tools.md): typed actions that agents can call. - [Prompts](/docs/build/prompts.md): prompt text versioned and runnable from code. - [Sandboxes](/docs/build/sandboxes.md), [Skills](/docs/build/skills.md), and [MCP](/docs/build/mcp.md): isolated workspaces, on-demand instructions, and external tools for agents. - SDK references: [Python](/sdk/python.md), [TypeScript](/sdk/typescript.md), and [Go](/sdk/go.md). Outcome: you can choose the smallest AGNT5 primitive for the behavior you need. ## 3. Build locally Run the worker with hot reload and trigger runs from Studio, the CLI, or the API before you deploy. - [Local development](/docs/build/local-development.md): run `agnt5 dev`, trigger local runs, and watch traces. - [Human in the Loop](/docs/build/human-in-the-loop.md): pause workflows for approvals, selections, and follow-up input. - [Webhooks](/docs/build/webhooks.md): receive signed external events and route them into workflows. - [Prompt caching](/docs/build/prompt-caching.md): reuse stable prompt prefixes to cut latency and token spend. - [Integrations](/docs/integrations/overview.md): store credentials once and scope them to a workspace, project, or environment. Outcome: your project runs locally with the inputs, prompts, approvals, events, and credentials it needs. ## 4. Go live A **deployment** is your worker running in AGNT5 Cloud with environment-scoped configuration and observable runs. - [Deploying](/docs/run/deploying.md): set secrets, deploy a worker, verify it, and trigger a run. - [Environments](/docs/run/environments.md): separate preview, staging, and production, and promote a verified deployment forward. - [Serverless endpoints](/docs/run/serverless-endpoints.md): run components on Cloudflare Workers, Vercel, Cloud Run, or another HTTP host. - [Event sources](/docs/integrations/event-sources/overview.md): connect Sentry, Slack, or another webhook publisher. Outcome: your workflow runs in a managed environment and receives production inputs. ## 5. Operate with trust A **run** is one execution; its trace, logs, metrics, and cost are the evidence you use to operate it. - [Runs](/docs/run/deploying.md#runs): list runs, describe one run, and open it in Studio. - [Traces](/docs/run/deploying.md#traces): inspect model calls, tool calls, inputs, outputs, retries, and timing. - [Logs](/docs/run/deploying.md#logs): stream deployment logs or fetch logs scoped to a run. - [Metrics](/docs/run/deploying.md#metrics): monitor executions, success rate, latency, and LLM cost. - [Streaming output](/docs/run/streaming-output.md): show live output while a run is in progress. Outcome: you can answer what happened during a production run and where to look next. ## 6. Improve behavior Decide whether a prompt, model, or workflow change is better than the current version, using real runs. - [Datasets](/docs/improve/datasets.md): collect representative inputs from production runs and publish a version. - [Scorers](/docs/improve/scorers.md): define how AGNT5 decides whether an output meets the target behavior. - [Experiments](/docs/improve/experiments.md): compare a candidate prompt, model, or agent against a baseline. - [Online evals](/docs/improve/online-evals.md): score sampled production runs and alert when quality drops. - [Prompts](/docs/build/prompts.md): version prompt changes and select versions from code. Outcome: you can turn observed behavior into safer prompt, model, and workflow changes. ## Next steps Pick up the path where you are. - [Run the quickstart](/docs/quickstart): Create and run the first workflow in the path. - [Deploy a worker](/docs/run/deploying): Move a working project into an AGNT5 environment. - [Measure quality](/docs/improve/overview): Turn production runs into datasets and experiments.