Knowledge and workflow platform
How to use
FastGPT.
Building grounded Q&A applications from documents, composing visual AI workflows and plugins, and publishing controlled agents through web channels or APIs.
FastGPT is a Open-source AI application platform for knowledge bases, conversational agents, visual workflows, plugins, Agent V2, and OpenAI-compatible APIs. Building grounded Q&A applications from documents, composing visual AI workflows and plugins, and publishing controlled agents through web channels or APIs. This guide covers the whole path in one place: official access, a first session that produces something reviewable, the checks that make output trustworthy, and the permissions worth limiting before you connect real work.
AI should make the work easier to inspect. If the workflow removes the source, the owner, or the review step, redesign the workflow.
Access & setup
Find, install, and sign in to FastGPT
Get into the official FastGPT experience with the right account and a setup you understand.
- 01
Start at https://fastgpt.io/ and confirm the domain before entering account or payment information.
- 02
Availability: FastGPT is used through a hosted or self-hosted web application and API rather than a standalone desktop client.
- 03
A FastGPT cloud account or secured self-hosted stack, configured chat, embedding and optional reranking models, team and key permissions, approved documents, object and vector storage, backups, monitoring, and cost and workflow limits.
- 04
Sign in with the account you intend to keep using, then review plan, data, notification, and permission settings.
- 05
Run one low-risk test task before connecting sensitive files, repositories, or workspace data.
- Use official download pages.
- Review permissions during setup.
- Keep installers and applications updated.
First session
Your first useful FastGPT session
Learn the interaction loop using a small task with a clear outcome.
- 01
Define the users, team, use case, source owners, data classes, answer and citation rules, model set, workflow actions, retention, privacy, quality, latency, budget, and human-review gates.
- 02
Create an isolated test team, configure scoped model credentials, build a small knowledge base from approved documents, and record document owner, version, rights, and expiry.
- 03
Create a simple conversational app before adding workflow nodes, require retrieved evidence and abstention, bound history and output, and keep actions read-only until independently authorized.
- 04
Test retrieval, citations, no-answer behavior, injection, permissions, all workflow branches, failures, token and concurrency cost, then publish with scoped expiring quotas and monitoring.
- State the outcome before the background.
- Provide the real source material.
- Review the result before expanding the task.
Quality control
Check the quality of FastGPT output
Establish that an autonomous run did the right thing, not merely that it finished.
- 01
Define what the run should achieve and what it must never touch before granting it a single tool.
- 02
Read the full execution trace: which tools were called, with what arguments, and in what order.
- 03
Verify the side effects directly in the target system rather than trusting the agent's own report of success.
- 04
Confirm failures surfaced as failures — a silent retry loop or a swallowed error is more dangerous than a crash.
- 05
Re-run the same task and compare: an agent that behaves differently across identical runs is not yet production-ready.
- Verify side effects in the system of record, not in the agent's summary.
- Require human approval for any irreversible or outward-facing action.
- Log every tool call so a run can be reconstructed afterwards.
Privacy & permissions
Use FastGPT safely
Bound what an autonomous system can reach before you let it run unattended.
- 01
Enumerate every tool, credential, and system the agent can reach, and remove the ones it does not need.
- 02
Require explicit human approval for irreversible actions: sending, publishing, paying, deleting, or deploying.
- 03
Run against non-production data until behaviour is predictable across repeated runs.
- 04
Set hard limits on spend, iterations, and runtime so a failure loop cannot run unbounded.
- 05
Treat anything the agent reads from the web or a document as data, never as instructions it may follow.
- FastGPT API keys act with the current member's access inside a team, so they may reach more than one app. OpenAI compatibility does not narrow that scope. Self-hosting requires secure cookies, origins, proxies, SSRF controls, databases, object and vector storage, code sandboxing, secrets, and patches. Isolate teams, minimize keys, validate uploads, constrain HTTP and tool nodes, require human approval, and test backup, restore, deletion, and billing.
- Follow your organisation's approved-use policy.
- Never treat fluent output as authorization to act.
Core workflows
Step-by-step ways to use FastGPT for the work it does best.
Each workflow is a separate guide with its own steps and review checkpoints.
Turn owned documents into a permissioned knowledge application with measurable retrieval, citations, abstention, freshness, and deletion.
↗ 02 WorkflowPublish a secure FastGPT workflow APIConvert a tested visual flow into a bounded API with scoped member credentials, explicit branches, safe nodes, quotas, and human release gates.
↗Official references
Check the current product documentation.
Features, plan limits, availability, and data controls change. These official pages are the starting points used for this guide.
- FastGPT overview ↗
- FastGPT quick start ↗
- FastGPT OpenAPI introduction ↗
- FastGPT environment and security settings ↗