Model guidesOfficial documentation7 min readModel guides

Learn FastGPT Plugin Core Technical Concepts

This reference page outlines core technical concepts for FastGPT plugins, tailored for engineers and technical decision makers evaluating or deploying Fas…

This reference page outlines core technical concepts for FastGPT plugins, tailored for engineers and technical decision makers evaluating or deploying FastGPT plugin-enabled systems.

Core Plugin Terminology Table

The following table defines standard FastGPT plugin terms:

TermDescription
PluginAn independent, reusable feature component. Plugins can have different types, such as tools, model presets, and Dataset sources.
Plugin PackageThe packaged .pkg file for a plugin. All plugin types are installed, updated, and managed through plugin packages.
ToolA plugin type that usually wraps third-party services, internal APIs, or local computation and can be called by workflows and Agents.
ToolkitA plugin that exposes multiple related child tools while sharing plugin metadata and secret configuration.
Plugin MarketplaceA centralized platform where users can search, download, and install plugins.
RuntimeThe backend implementation responsible for executing plugin code. The current default runtime is local-pool.
PodA single plugin child process in the local process pool. One plugin service can own multiple Pods.

Plugin Type Functional Roles

Plugins are categorized by their intended use case within the FastGPT ecosystem:

  • Tool plugins: Wrap third-party services, internal APIs, or local computational tasks, and are callable directly by FastGPT workflows and AI Agents.
  • Toolkit plugins: Aggregate multiple related tool plugins, sharing unified plugin metadata and secret configuration across all included child tools to reduce redundant setup.
  • Specialized plugin types: Including model presets and Dataset source plugins, which extend core FastGPT functionality beyond external service integration.

Runtime and Pod Architecture

The FastGPT plugin runtime provides the backend execution layer for all plugin code. The default runtime environment is local-pool, which manages a pool of isolated child processes called Pods. Each Pod represents a single running instance of a plugin service, and a single plugin service can deploy multiple Pods to support varying levels of concurrent plugin invocation demand.

Source: FastGPT official source

Applicability and version scope

Use this page for the documented Model guides scenario. Confirm the FastGPT, dependency, API, and deployment versions in the official source before applying a change.

Safety guardrails

Use [REDACTED_CREDENTIAL] for credentials and private data. Confirm the documented environment and version before review.

Rollback guidance

Restore the prior technical-content authority snapshot. Restore saved configuration and data snapshots, then repeat the smallest verification scenario.