Tool Catalog

For creating your own tools, see Custom Tools.

Quick Start

// Register tools with an agent
Agent agent = AgentBuilder.create()
    .llm(LLMClientFactory.create("openai", "gpt-4o", 0.7f))
    .role(myRole)
    .tool(new BraveSearchTool())
    .tool(new CalculatorTool())
    .tool(new WikipediaTool())
    .build();

// The LLM will call tools automatically during conversation
String response = agent.chat("What is the population of Tokyo? Calculate 3x that number.");

Tools require API keys via environment variables:

export BRAVE_API_KEY=your-key
export SERPAPI_API_KEY=your-key

Search (18 tools)

Search tools let your agent find information on the web, in knowledge bases, and across package registries. Some require an API key (set as an environment variable), while others work out of the box.

Academic (19 tools)

Academic tools give your agent access to scholarly papers, preprints, citation data, and researcher profiles. They are useful for research assistants, literature review agents, and scientific Q&A.

File (10 tools)

File tools let your agent read, write, and transform files on the local filesystem. This includes plain text, PDFs, CSV, JSON, and XML formats.

Communication (7 tools)

Communication tools allow your agent to send emails, post messages to Slack or Discord, and send SMS. Each tool requires credentials for the corresponding service.

FinTech (8 tools)

FinTech tools enable your agent to process payments, create invoices, and manage product catalogs through services like Square, Cash App, and Lightning Network.

Database (4 tools)

Database tools let your agent query SQL databases, MongoDB collections, Redis key-value stores, and vector databases directly. Each tool needs a connection string or URL configured as an environment variable.

Developer (4 tools)

Developer tools help your agent interact with code repositories, execute Java snippets, analyze project structure, and manage dependencies.

Productivity (5 tools)

Productivity tools connect your agent to popular project management and workspace platforms like Google Calendar, Google Drive, Notion, Trello, and Jira.

Media (5 tools)

Media tools handle audio processing, including speech-to-text transcription and text-to-speech synthesis. They are useful for building voice-enabled agents.

Social (4 tools)

Social tools let your agent interact with social media platforms -- posting tweets, browsing Reddit, and retrieving LinkedIn data.

Utility (5 tools)

Utility tools provide common helper functions that agents frequently need, such as math calculations, date/time operations, hashing, QR codes, and translation.

Diagram (3 tools)

Diagram tools let your agent generate visual diagrams and architecture charts programmatically, without needing a manual drawing tool.

Code Execution (3 tools)

Code execution tools let your agent run code snippets in sandboxed environments. This is useful for agents that need to compute results, test code, or process data dynamically.

Real-time Data (3 tools)

Real-time data tools provide your agent with live information such as cryptocurrency prices, weather conditions, and currency exchange rates.

Visualization (3 tools)

Visualization tools let your agent produce charts, data tables, and infographics from structured data, turning raw numbers into visual outputs.

Agentic (4 tools)

Agentic tools are meta-level tools that enable higher-order agent behaviors -- delegating work to sub-agents, summarizing documents intelligently, and conducting multi-step research.

AI / Self-Evolution (1 tool)

The self-evolution tool allows an agent to learn from past interactions by recording feedback, optimizing its prompts, and extracting reusable skills -- all without external services.

SelfEvolutionTool provides nine operations for prompt optimization and agent self-improvement, based on research in self-evolving agents (EvoAgentX, TextGrad, DSPy).

Data (feedback, skills, performance) is persisted to ~/.tnsai/evolution/. No external API key required.

Agent agent = AgentBuilder.create()
    .llm(client)
    .role(myRole)
    .tool(new SelfEvolutionTool())
    .build();

// The agent can now record feedback and optimize its own prompts
agent.chat("{\"operation\": \"record_feedback\", \"prompt_id\": \"research\", \"success\": true, \"rating\": 4}");
agent.chat("{\"operation\": \"optimize_prompt\", \"prompt_id\": \"research\", \"prompt\": \"Research the topic\"}");

Document (5 tools)

Document tools handle reading, converting, and generating documents. They can extract text from URLs and files, convert between formats, and create slide presentations.

DocumentReaderTool

Extracts plain text from URLs and local files. Supports HTML (strips tags, extracts body text), Markdown (renders to text), and PDF (text extraction).

Agent agent = AgentBuilder.create()
    .llm(client)
    .tool(new DocumentReaderTool())
    .build();

// Read from URL
agent.chat("Read and summarize https://example.com/article");

// Read local file
agent.chat("Extract text from /path/to/report.pdf");

The tool accepts a source parameter (URL or file path) and returns the extracted text content. For HTML pages, it strips scripts, styles, and navigation elements to return clean body text.

Additional Categories

Beyond the main categories above, TnsAI includes specialized tool groups for e-commerce, system integration, scraping, AI vision, finance, CRM, knowledge graphs, and external memory.