Deployment and upgradesOfficial documentation5 min readDeployment and upgrades

Add Custom SiliconCloud Models to FastGPT

Default and Custom SiliconCloud Model Setup The self-hosted FastGPT system includes a small set of pre-configured SiliconCloud models for quick initial te…

Default and Custom SiliconCloud Model Setup

The self-hosted FastGPT system includes a small set of pre-configured SiliconCloud models for quick initial testing. To add additional models beyond this default collection, users must follow the official manual custom model addition process, linked at ./intro.en.mdx#add-a-custom-model.

Example Target Model Configuration

This demonstration uses a standard set of SiliconCloud models to enable a complete AI workflow. The configured models cover core inference, embedding, reranking, text-to-speech (TTS), and speech-to-text (STT) functionality:

Model TypeModel IdentifierPrimary Use Case
Text & Vision ModelQwen2.5 72bDual text generation and visual understanding
Embedding Modelbge-m3Vector embedding generation for semantic search
Rerank Modelbge-reranker-v2-m3Reorder search results to improve relevance
TTS Modelfish-speech-1.5Convert written text to natural speech audio
STT ModelSenseVoiceSmallConvert spoken audio to written text transcript

Configuration Validation

After completing the manual model addition steps outlined in the linked guide, users can cross-reference their setup with the included reference screenshot (image-104.png). This visual resource confirms the correct formatting and placement of all configured SiliconCloud model entries within the FastGPT administrative interface.

Source: FastGPT official source

Applicability and version scope

Use this page for the documented Deployment and upgrades 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.