// MCP TOOL REFERENCE

PRAISONAI TOOLS

91 tools from the PraisonAI MCP Server, categorised by risk level.

// ALL 91 PRAISONAI TOOLS

READ 74 tools

Read --criteria Custom criteria for validation Read --model Model to use (provider/model format) Read --quality Image quality (standard, hd) Read --size Image size (1024x1024, 1792x1024, 1024x1792) Read --store Vector store provider (memory, pinecone, etc.) Read --stream Enable streaming output Read --style Image style (vivid, natural) Read --timeout Request timeout in ms Read --user-id User ID for memory isolation Read --verbose Show detailed output Read adapters List available adapters Read airweaveSearch Search across connected data sources using semantic search. Returns relevant context for RAG applications. Read blocked_words Check for blocked words Read chroma ChromaDB vector database Read ChromaVectorStore ChromaDB embedding database Read cohere Cohere reranking API Read CohereReranker Cohere reranking API Read config Show resolved configuration Read ConsoleObservabilityProvider Console-based logging Read contains_keywords Output contains: ${keywords.join( Read cross-encoder Cross-encoder model Read CrossEncoderReranker Cross-encoder model reranking Read display-types List display callback types Read events List hook events (alias for list) Read file File-based persistent cache Read firecrawlCrawl Crawl a website and extract content from multiple pages. Read firecrawlScrape Scrape a single web page and extract its content in LLM-ready format. Read guard Check text for security threats like prompt injection, jailbreaks, and data exfiltration attempts. Read help Show this help Read info Show database information Read json Demo structured output Read LangfuseObservabilityProvider Langfuse integration Read length Output length between ${minLength}-${maxLength} Read list List all cached entries Read llm LLM-based reranking Read LLMReranker LLM-based reranking Read map Show repository map Read max_length Ensure content is under ${maxChars} characters Read memory In-memory vector store (default) Read MemoryObservabilityProvider In-memory trace storage Read MemoryVectorStore In-memory vector store for development Read min_length Ensure content is at least ${minChars} characters Read models List common models by provider Read no_harmful Output contains no harmful content Read openai OpenAI TTS and Whisper Read optimize Optimize context Read parallelSearch Search the web using Parallel with content extraction capabilities. Read pattern Check content against pattern: ${regex} Read pending List pending approval requests Read perplexitySearch Search the web using Perplexity for real-time information and answers. Read pinecone Pinecone vector database Read PineconeVectorStore Pinecone cloud vector database Read postgres PostgreSQL (Neon) Read providers List available AI SDK providers Read qdrant Qdrant vector database Read QdrantVectorStore Qdrant vector database Read query_database Query the PostgreSQL database using natural language. Available tables: ${client.getSchema().map(t => t.nam... Read redis Redis (Upstash) Read relevance Output is relevant to input Read required_words Check for required words Read sqlite SQLite database Read stats Show cache statistics Read status Show approval status Read tavilyCrawl Crawl a website using Tavily. Discovers and extracts content from multiple pages starting from a base URL. Read tavilyExtract Extract structured content from web pages using Tavily. Provide URLs to extract their content. Read tavilySearch Search the web for information using Tavily. Returns relevant results with content snippets and optional AI... Read tools Demo tool calling Read trace Demo attribution/trace headers Read valid_json Ensure content is valid JSON Read verify Verify claims against source materials. Check if statements are supported by provided evidence. Read voices List available voices Read weaviate Weaviate vector database Read WeaviateVectorStore Weaviate vector search engine Read webSearch Search the web using Exa AI-powered search. Returns semantically relevant results with content.

WRITE 8 tools

Write bedrockBrowserClick Click an element on the page using a CSS selector. Write bedrockBrowserFill Fill a form field with text using a CSS selector. Write budget Manage token budget Write commit Commit pending changes Write create Create context manager Write disable Disable telemetry collection Write enable Enable telemetry collection Write export Export telemetry data

DESTRUCTIVE 3 tools

Destructive cancel-all Cancel all pending requests Destructive clear Clear all cache entries Destructive redact Remove sensitive data (PII, PHI, secrets) from text. Redacts SSNs, emails, phone numbers, credit cards, etc.

EXECUTE 6 tools

Execute bedrockBrowserNavigate Navigate to a URL using cloud-based browser automation. Execute bedrockCodeInterpreter Execute code in a secure AWS-managed sandbox. Supports Python, JavaScript, and TypeScript. Execute build Build context string Execute codeMode Execute code in a sandboxed environment with access to imported tools. Write files, run code, and get results. Execute executeCode Execute Python code in a secure sandbox environment. Use this for calculations, data processing, and comput... Execute interactive Start interactive approval mode

// RUNNING THIS SERVER

The managed route: connect PraisonAI through the PolicyLayer gateway — every tool call above is checked against your policy before it runs, with a full audit log.

DIRECT INSTALL (UNMANAGED) npx -y pypi:praisonai

// FAQ

How many tools does the PraisonAI MCP server have? +

The PraisonAI MCP server exposes 91 tools across 4 categories: Read, Write, Destructive, Execute.

How do I enforce policies on PraisonAI tools? +

Route the PraisonAI server through the PolicyLayer gateway. Define allow, deny, or approval rules per tool in the dashboard — they are enforced on every call before it reaches the server.

What risk categories do PraisonAI tools fall into? +

PraisonAI tools are categorised as Read (74), Write (8), Destructive (3), Execute (6). Each category has a recommended default policy.

Let agents act without letting them run wild.

Route your MCP servers through PolicyLayer and every tool call is checked against your policy before it runs — allow, deny, or require approval. Per-identity grants. Full audit log. Live in minutes.

SECURE YOUR MCP →

Free to start. No card required.

4,600+ MCP servers and 31,000+ tools scanned and risk-classified.