FastGPT AI Agent Platform
A knowledge-based AI agent platform on LLMs with out-of-the-box data processing, RAG retrieval, and visual workflow orchestration.
- Source repo
- labring/FastGPT
- Stars
- ★ 30k
- Last updated
- today
- License
- NOASSERTION
- Primary language
- TypeScript
- FA score
- 0/100 · Major gaps
At a glance
- How it runs
- Works with
- Universal · cross-platform
- Cost
- Free tier plus a paid hosted plan
- Setup effort
- Medium · a few setup steps
- You'll need
- Typical use
- Enterprise internal knowledge base: Import employee handbooks, training docs, PDFs, and let staff query company policies and procedures conversationally.
- Not a fit if
- Teams that need fully offline operation without external LLM APIs
- Users expecting all advanced features and deep support for free
- Teams wanting to offer SaaS services despite the license ban
- Source review
- 0/100 · Major gaps 8 safety controls not found
What does this agent do, and when should you use it?
FastGPT is an AI agent building platform developed by labring, designed to create knowledge-based question-answering systems powered by large language models. It offers complete data processing capabilities, supporting imports from TXT, MD, HTML, PDF, Docx, PPTX, CSV, and XLSX files, as well as URL reading and CSV batch import. Core features include multi-knowledge-base reuse, hybrid retrieval with re-ranking, chunk-level editing and deletion, and API-based knowledge bases. Users can orchestrate complex workflows using a visual Flow editor, which supports Agent Skill orchestration, user interaction nodes, bidirectional MCP, and basic RPA nodes. The platform provides debugging capabilities such as single-point knowledge base search testing, citation feedback, complete call-chain logs, and application evaluation. Operations features include no-login sharing, iframe embedding, and conversation annotation.
FastGPT reads various document formats (TXT, MD, HTML, PDF, Docx, PPTX, CSV, XLSX) and URLs, performs text splitting and QA import to build knowledge bases. It retrieves information using hybrid search and re-ranking algorithms, supporting multi-base reuse. Users can create conversational and plugin workflows via a visual Flow editor, orchestrating model calls, user interactions, and MCP integrations. The platform executes workflows by invoking LLM APIs to generate responses with citation sources. It also offers app debugging and operations features like search testing, log viewing, conversation annotation, and iframe embedding. Deployment is done via Docker and docker compose, with a Next.js-based frontend.
- Enterprise internal knowledge base: Import employee handbooks, training docs, PDFs, and let staff query company policies and procedures conversationally.
- Customer support automation: Build a RAG bot on product documentation to auto-answer common queries, with human annotation and refinement.
- Research literature management: Upload papers and reports, quickly locate relevant passages via hybrid search for lit reviews and citations.
- Educational tutoring: Teachers upload textbooks and lecture notes to create a Q&A system for personalized student support.
- Legal/compliance document review: Import laws and compliance handbooks, query with citations to assist review work.
- Developer tool: Load API docs and code examples, get quick technical answers through conversation.
How do you install or deploy this agent?
Quick deployment via Docker: run bash <(curl -fsSL https://doc.fastgpt.io/deploy/install.sh) to pull the config, then docker compose up -d to start. Access at http://localhost:3000 with default credentials root / 1234.
How do you use this agent?
After accessing the FastGPT UI, create a knowledge base and upload documents (supports TXT, MD, HTML, PDF, Docx, PPTX, CSV, XLSX, or URLs); the system splits the text automatically. In the app editor, select the knowledge base, and use the Flow visual editor to orchestrate workflows with model calls and user interaction nodes. Once configured, test and deploy the chatbot in the conversation interface, embed it via iframe, or share a no-login link.
What are this agent's strengths and limitations?
- Out-of-the-box data processing: supports multiple file formats and URL import, automated splitting and QA import, reducing setup effort.
- Visual workflow orchestration: create complex conversational and plugin workflows without extensive coding.
- Multi-base reuse and hybrid retrieval: mix multiple knowledge bases with hybrid search and re-ranking for better accuracy.
- Comprehensive debugging and operations: search testing, call-chain logs, evaluation, annotation, etc., aid optimization.
- License restrictions: not allowed to provide SaaS services, commercial use requires keeping copyright info and may need commercial licensing.
- Deployment requires Docker, which can be a barrier for non-technical users.
- Some advanced features require the commercial version; community edition lacks full support and deep services.
- Incomplete features: e.g., advanced workflow debugging mode and RAG module hot-reload are not yet implemented.
- External LLM API dependency: you need to bring your own model API or use defaults, incurring potential extra costs.
How does this agent compare with similar options?
Compared to other open-source LLM application platforms like Dify, FastGPT focuses more on knowledge-base Q&A scenarios with deeper data processing and RAG optimization, but may have a smaller ecosystem and fewer plugins.
Key facts side by side with the most closely related agents.
| Agent | Source review | Form / cost | Stars | Updated | Language | Full support on |
|---|---|---|---|---|---|---|
| FastGPT AI Agent Platform This agent | 0 · Major gaps | Self-hosted serviceFreemium | ★ 30k | today | TypeScript | — |
| CAMEL Multi-Agent Framework | 61 · Some gaps | Library / SDKFree + model costs | ★ 18k | 7d ago | Python | OpenAI API |
| DB-GPT: Open-Source Agentic AI Data Assistant | 54 · Major gaps | CLIFree + model costs | ★ 20k | 8d ago | Python | — |
| Graphiti: Temporal Context Graph Engine for AI Agents | 50 · Major gaps | Library / SDKFree + model costs | ★ 32k | today | Python | OpenAI API · Claude API |
How does FollowAgents rate this agent?
Why each dimension lost points
Insufficient evidence: no concrete implementation of permission model, user confirmation, data flow transparency, sensitive data handling, dependency security, external effects, rollback, or source attribution. Hence all trust criteria scored 0.
Insufficient evidence: no concrete implementation of self-consistency, dependency availability, or failure messages. Hence all reliability criteria scored 0.
Insufficient evidence: no concrete implementation of audience, capability boundaries, trigger precision, or environment fit. Hence all adaptability criteria scored 0.
Insufficient evidence: no concrete implementation of information architecture, install notes, naming stability, examples and FAQ, known limitations, license, versioning changelog, or maintenance responsibility. Hence all convention criteria scored 0.
Insufficient evidence: no concrete implementation of output usability, marginal value, or cost benefit. Hence all effectiveness criteria scored 0.
Insufficient evidence: no concrete implementation of claim traceability, cross-source corroboration, or fact-inference separation. Hence all verifiability criteria scored 0.
- Not found in source: least-privilege scopingGrant only what the task needs: a dedicated account or read-only token, scoped to specific directories and repos.
- Not found in source: confirmation before actingTurn on (or add) a confirmation step before it acts, and try it in a sandbox or test environment before real data.
- Not found in source: data-flow disclosureWatch which external services it contacts (proxy or firewall logs) and keep sensitive data out until you know where it goes.
- Not found in source: sensitive-data handlingUse dedicated, low-privilege, revocable API keys — never production credentials — and keep secrets out of logs.
- Not found in source: dependency securityPin versions and run a dependency audit (npm audit, pip-audit) before installing; prefer running it in a container.
- Not found in source: disclosed external effectsEstablish which external systems it writes to, sends to or changes, and verify with test accounts or repos before production.
- Not found in source: rollback or recovery pathBack up first, or work on a git branch or snapshot, so its changes can be undone.
- Not found in source: verifiable attributionInstall from the official repo or registry and check the publisher and URL to avoid look-alike packages.
- The provided files are insufficient for a full FARS-2.1 assessment; only README, license, security policy, package.json, workflows, and test files are included.
- No AGENTS.md manifest is provided, but it is not required.
- Publisher identity is unverified, but no deduction is made for that alone.
Recent releases
Latest entries from the project's GitHub Releases. Notes are quoted from upstream as written, not rewritten or tested by FollowAgents.
-
Release notes excerpt
📦 升级指南 镜像更新 - 更新 fastgpt-app（FastGPT 主服务）镜像 tag: v4.17.2 - 更新 fastgpt-pro（FastGPT 商业版）镜像 tag: v4.17.2 升级脚本 本次升级包含以下手动触发升级脚本（可在「管理员 > 版本升级」中查看与执行）： 1. 历史分块训练任务迁移（20261008_migrate_chunk_training）：将存量旧版 chunk 训练任务平滑转换为关联正式数据行的新模式。属于手动触发任务（不阻塞服务启动），请在确认集群中所有旧版本 FastGPT App 与 Pro 实例已完全停止、旧任务租约过期后再点击执行。 2. 知识库重建状态迁移（20261009_migrate_dataset_rebuild_status）：依赖上述训练任务迁移完成，属于手动触发任务（不阻塞服务启动），将历史知识库数据行中的旧版 rebuilding 标记和过渡状态统一迁移至全新的 indexStatus 重建状态体系。 🚀 新增内容 1. 知识库数据提前落库与训练链路重构： …View on GitHub ↗ -
Release notes excerpt
📦 升级指南 镜像更新 - 更新 fastgpt-app（FastGPT 主服务）镜像 tag: v4.17.1 - 更新 fastgpt-pro（FastGPT 商业版）镜像 tag: v4.17.1 - 更新 fastgpt-plugin 镜像 tag: v1.1.4 - 更新 AI Proxy 镜像 tag: v0.7.3 升级脚本 本次升级包含以下自动化脚本 1. 回填应用资源权限快照，清理遗留 V1 应用及版本数据，并校验应用正式版本；迁移任务不阻塞服务启动，可在系统升级任务中查看进度和失败记录。 2. 为兼容渠道启用 reasoning 映射：为现有 AIProxy OpenAI 协议渠道开启 map_reasoning_to_reasoning_content: true 的响应字段转换，解决深度思考模型（如 DeepSeek-R1）的思考内容字段兼容问题，并保留渠道的原有配置；已配置该项的渠道不会重复更新，迁移任务不阻塞服务启动。 3. 升级知识库标签数据：将存量 Collection 标签迁移为类型化标签，并清理重复标签定义；迁移任务不阻塞服务启动。 …View on GitHub ↗ -
Release notes excerpt
升级说明 升级到 V4.17.0 前，请先升级到 V4.16.2，并执行完成 V4.16.x 版本要求的所有旧升级脚本。 V4.17.0 的自动升级机制只管理从本版本开始登记的新任务，不会补执行此前版本的历史脚本。 从 4.17.0 版本开始，登记到系统升级机制的任务会自动执行，并通过 lease 抢占避免多节点并发问题，同时支持心跳检测，检查脚本是否活跃。可在 管理员 > 版本升级 中查看升级脚本状态。 分批迁移任务默认每批处理 100 条记录。数据量较大时，可通过 SYSTEM_MIGRATION_BATCH_SIZE 调整为 50～1000。 AI Proxy（必需） 从 V4.17.0 起，必须部署或接入 AI Proxy，并在 FastGPT App 及复用 App/Service 配置的商业版 Admin 服务中设置： - AIPROXY_API_ENDPOINT：可访问的 AI Proxy 服务根地址，例如 http://aiproxy:3000，不要附加 /v1。 - AIPROXY_API_TOKEN：与 AI Proxy 服务端配置一致的非空管理员 Token。 …View on GitHub ↗