Read text file

Rebuild the index without deleted items.

Redistribute flattened codes back into SNAC's 3-layer format. Terminates early if invalid codes are detected (out of range 0

Register default tools from local_llm_serving

Register default tools

Create a rule-based summary when LLM is not available. Args: question: The original question respons

Run one iteration with Anthropic API

Run a multi-session test

Run one iteration with OpenAI/OpenRouter API

Internal function to run a recurring timer.

Run a single session test

Run task with browser-use agent and capture workflow. Args: max_steps: Maximum steps for agent

Run task using a learned workflow. Args: workflow: The workflow to replay Returns:

Save experiences to file.

Save index to disk.

Save BM25 indexes to disk

Save the captured workflow to knowledge base.

Save results to JSON file

Save benchmark results

Save current trajectory to JSON file for debugging

Save current trajectory to JSON file for debugging

Save current trajectory to file for debugging

Perform BM25 search. Educational Note: BM25 excels at finding exact term matches and handles technical terms

Search and return matching lines with context

Count matches per file

Search using Dify API

Search using Dify API

Perform semantic search using embeddings. Educational Note: Embedding search captures semantic meaning and relations

Search and return files that have matches

Search using GraphRAG knowledge graph index

Search using GraphRAG knowledge graph index

Search using local retrieval pipeline

Search using local retrieval pipeline

Search using RAPTOR tree-based index

Search using RAPTOR tree-based index

Send email using SendGrid API.

Send email using SMTP.

Split text into sentences

Split text into sentences (simple implementation)

Split text into sentences (simple implementation)

Store full document locally

Store full document locally

Stream response from Gemini with thinking mode for debugging

Stream response from OpenAI-compatible API

Stream response content

Stream response content

Pick up an item.

Add a new memory

Execute Python code in restricted environment

Delete a memory

Execute shell command

Read file contents with optional line-based reading

Rewrite TODO list with new pending items

Update an existing memory

Update status of TODO items

Write content to file

Truncate long values for display. Args: value: Value to truncate max_length: Maximum length

Try to craft items (hidden mechanic).

Update index statistics

Use an item.

Wait for admin to respond to approval request.

Wait for server to be ready

Add a document to the store.

Add a document to the local store

Add a document to the store. Args: text: Document text doc_id: Optional document ID (will be generat

Add a document to the local store

Add an item to the ANNOY index.

Add a new metric to the collection

Add a step to the chain Args: request: Request for this step **kwargs: Additional parameters for pro

Analyze the impact of each ablation factor

Analyze audio with a specific query

Analyze image using AI (OpenAI Vision API). Args: image_path: Path to image file prompt: Prompt for AI analysis

Analyze a PDF document with a specific query

Analyze ablation study results Args: results: List of test results Returns: Ana

Analyze video content using AI vision on keyframes. Args: video_path: Path to video file num_frames: Number of frames to

Apply memory updates to the memory manager Args: updates: List of memory updates to apply R

批量搜索多个问题 Args: questions: 问题列表 Returns: 答案列表

Benchmark the basic Elo implementation.

Benchmark the NumPy + Numba optimized implementation.

Execute a high-level browser task using the browser-use agent. Args: task: Natural language description of the task to perform

Get content from the current page. Args: selector: Optional CSS selector to extract specific content Returns: Di

List all open browser tabs. Returns: Dictionary with list of tabs

Navigate to a URL in the browser. Args: url: The URL to navigate to new_tab: Whether to open in a new tab Return

Take a screenshot of the current page. Args: full_page: Whether to capture the full page or just viewport Returns:

* Convert Buffer to Float32Array for Silero VAD * @param {Buffer} buffer - PCM 16-bit audio buffer * @returns {Float32Array} Normalized audio sa

Build confusion matrix and identify problematic languages. Returns: Dict with confusion matrix, per-language stats, and error ex

Build and save the complete dataset

Build and save the complete dataset

Build RAPTOR and/or GraphRAG index from text or file.

Build RAPTOR and/or GraphRAG indexes from a document.

Build Elo leaderboard from voting data. Args: df: DataFrame with columns 'model_a', 'model_b', 'winner' initial_rating: