Rerank chunks for each query using Liquid AI's ColBERT model. This is useful for reranking results from a first-stage retriever (lik

Return number of entries in current collection.

Return number of entries in current collection.

Text completion using the chat endpoint. Args: prompts: Single prompt or list of prompts **kwargs: A

Generate completions for the given text prompts. Args: prompts: String or list of strings to generate completions for

Compute FLOPs, MACs, and parameters for transformer models using the calflops package. Args: model_name (str): HuggingFace model

Create a new File Search store for RAG. Args: display_name: Display name for the file search store

Create a comparison chart for the methods.

Use Exa's deep-reasoning for synthesized answers with grounding citations. Args: query: The research query. output_sc

Delete the current collection from ChromaDB. :returns Whether the collection was successfully deleted

Delete an uploaded file. Example: client.delete_file("file_abc123")

Delete a File Search store. Args: store_name: Name of the store to delete (format: fileSearchStores/xxxxx)

Divide a by b and return the result. Returns error if b is zero.

Generate embeddings using Vercel AI Gateway. Args: content: Text content to embed (single string or list of strings).

Generate embeddings using the Mistral embeddings API. Args: inputs: List of strings to embed **kwargs: Addit

Generate embeddings using Gemini's embedding API. See: https://ai.google.dev/gemini-api/docs/embeddings

Embedding not supported by Distributed Inference API.

Generate embeddings using the OpenRouter API.

Generate embeddings using the Perplexity Embeddings API. Args: content: Text or list of texts to embed. model: Embedd

Embed content using model (if supported).

Generate embeddings using SambaNova's embedding API. Args: content: Text content to embed (single string or list

Generate embeddings using the Osaurus server. Note: Embedding support depends on the specific model loaded in Osaurus.

Generate contextualized embeddings for document chunks. Args: documents: Nested list where each inner list is ordered chunks from

Generate embeddings for the given input text. Args: input_text: The text to generate embeddings for

Encode texts to create embeddings.

Encode texts to create embeddings using Gemini model. Args: texts: Single text or list of texts to encode ta

Encode texts to create embeddings using OpenRouter model. Args: texts: Single text or list of texts to encode

Encode texts to create embeddings using Ollama model. Args: texts: Single text or list of texts to encode **

Encode texts to create embeddings using Mistral model. Args: texts: Single text or list of texts to encode *

Encode texts to create embeddings using Modular MAX Serve. Args: texts: Single text or list of texts to encode

Encode texts using model by name (for backward compatibility)

Encode texts using model by name (for backward compatibility)

Encode texts using model by name (for backward compatibility)

Encode texts using model by name (for backward compatibility)

Encode texts using model by name (for backward compatibility)

Encode texts using model by name (for backward compatibility)

Encode texts using model by name (for backward compatibility)

Encrypt and sign a message using sequential processing. This function has the same API as encrypt_and_sign but processes large messages

Ensure metadata exists and has skill_id.

Generate SSE events.

Execute an MCP tool synchronously

Execute tool calls and return the results. Args: tool_calls: List of tool call objects from the LLM Returns: List of diction

Export comparison results to a CSV file.

Extract specific entity types from text and return as a simple list. Args: text: Input text to process entity_types: Lis

Extract content from a single URL and return as a dictionary. This is a convenience wrapper around parallel_extract for single URL extra

Format response text for display in Streamlit.

Format story data for display in the UI. Args: story_data: Story data dictionary Returns: Formatted story s

Create a ConversationHistory from a dictionary.

Generate audio from text and return the path to the audio file.

Generate a demo story without requiring API keys

Generate diffs between workspace states. Args: before_state: Previous workspace state after_state: C

Generate images using the MLX Omni Server. Args: prompt: The prompt to generate images from n: Number of ima

Generate images using the MLX Omni Server. Args: prompt: The prompt to generate images from n: Number of ima

Get the public agent card.

Get the list of file IDs currently attached to chat. Returns: List of file IDs

Get list of available Gemini embedding models.

Get list of available Mistral embedding models.

Get list of commonly used embedding models with Modular MAX.

Get a list of available models through Cloudflare AI Gateway. Note: Since Cloudflare AI Gateway proxies to multiple provider

Get a list of available Gemini models Returns: List[str]: List of model names

Get a list of available models from Parallel AI. Returns: List[str]: List of model names available through Parallel AI

Get list of available models from OpenRouter.

Get a list of available models from Novita AI. Returns: List[str]: List of model names available through Novita

Get a list of commonly supported models for LMCache. Returns: List[str]: List of model names that work well with

Get a list of available models from Perplexity AI. Returns: List[str]: List of model names available through Per

Get a list of available Grok models from the X.AI API. Returns: List[str]: List of model names available through X.AI

Get a list of available MLX models from the server. Returns: List[str]: List of model names

Get a list of available Ollama models Returns: List[str]: List of model names

Get a list of available models from Tencent Hunyuan. Returns: List[str]: List of model names available through H

Get a list of available Osaurus models. Note: This is a static method that returns common Osaurus model names. For d

Get a list of available MLX models from the server. Returns: List[str]: List of model names

Get a list of available Gemini models Returns: List[str]: List of model names

Get a list of available Ollama models Returns: List[str]: List of model names

Low-level method to get the raw chat completion response. Args: messages: List of message dictionaries with 'rol

Extract the system prompt from the character exposé for chat.

Get citations from the last chat response. Returns: List of citation objects from the last response, or None if

Get current configuration.

Check AI Gateway credit balance and usage information. Reference: https://vercel.com/docs/ai-gateway/usage#credits

Get the extended agent card with authentication.

Get the extended agent card for authenticated users.

Get information about a specific file. Example: file_info = client.get_file("file_abc123") print(f"Filename:

Retrieve detailed information about a specific generation by its ID. This endpoint allows you to look up usage data, costs, and metadata for

Get grounding metadata from the last response. This includes citations from File Search and other grounding sources.

Check the health of the server.

Get detailed usage information from the last Interactions API call. This provides more detailed usage information than the s