You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
4.8 KiB
4.8 KiB
Module Support
DeepSearcher supports various integration modules including embedding models, large language models, document loaders and vector databases.
📊 Overview
| Module Type | Count | Description |
|---|---|---|
| Embedding Models | 7+ | Text vectorization tools |
| Large Language Models | 11+ | Query processing and text generation |
| Document Loaders | 5+ | Parse and process documents in various formats |
| Vector Databases | 2+ | Store and retrieve vector data |
🔢 Embedding Models
Support for various embedding models to convert text into vector representations for semantic search.
| Provider | Required Environment Variables | Features |
|---|---|---|
| Open-source models | None | Locally runnable open-source models |
| OpenAI | OPENAI_API_KEY |
High-quality embeddings, easy to use |
| VoyageAI | VOYAGE_API_KEY |
Embeddings optimized for retrieval |
| Amazon Bedrock | AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY |
AWS integration, enterprise-grade |
| FastEmbed | None | Fast lightweight embeddings |
| PPIO | PPIO_API_KEY |
Flexible cloud embeddings |
| Novita AI | NOVITA_API_KEY |
Rich model selection |
| IBM watsonx.ai | WATSONX_APIKEY, WATSONX_URL, WATSONX_PROJECT_ID |
IBM's Enterprise AI platform |
🧠 Large Language Models
Support for various large language models (LLMs) to process queries and generate responses.
| Provider | Required Environment Variables | Features |
|---|---|---|
| OpenAI | OPENAI_API_KEY |
GPT model family |
| DeepSeek | DEEPSEEK_API_KEY |
Powerful reasoning capabilities |
| XAI Grok | XAI_API_KEY |
Real-time knowledge and humor |
| Anthropic Claude | ANTHROPIC_API_KEY |
Excellent long-context understanding |
| SiliconFlow | SILICONFLOW_API_KEY |
Enterprise inference service |
| PPIO | PPIO_API_KEY |
Diverse model support |
| TogetherAI | TOGETHER_API_KEY |
Wide range of open-source models |
| Google Gemini | GEMINI_API_KEY |
Google's multimodal models |
| SambaNova | SAMBANOVA_API_KEY |
High-performance AI platform |
| Ollama | None | Local LLM deployment |
| Novita AI | NOVITA_API_KEY |
Diverse AI services |
| IBM watsonx.ai | WATSONX_APIKEY, WATSONX_URL, WATSONX_PROJECT_ID |
IBM's Enterprise AI platform |
📄 Document Loader
Support for loading and processing documents from various sources.
Local File Loaders
| Loader | Supported Formats | Required Environment Variables |
|---|---|---|
| Built-in Loader | PDF, TXT, MD | None |
| Unstructured | Multiple document formats | UNSTRUCTURED_API_KEY, UNSTRUCTURED_URL (optional) |
Web Crawlers
| Crawler | Description | Required Environment Variables/Setup |
|---|---|---|
| FireCrawl | Crawler designed for AI applications | FIRECRAWL_API_KEY |
| Jina Reader | High-accuracy web content extraction | JINA_API_TOKEN |
| Crawl4AI | Browser automation crawler | Run crawl4ai-setup for first-time use |
💾 Vector Database Support
Support for various vector databases for efficient storage and retrieval of embeddings.
| Database | Description | Features |
|---|---|---|
| Milvus | Open-source vector database | High-performance, scalable |
| Zilliz Cloud | Managed Milvus service | Fully managed, maintenance-free |
| Qdrant | Vector similarity search engine | Simple, efficient |