AI agent tools

An AI agent is only as capable as the tools it can call. Elevence AI ships 12+ pre-built tools that any custom agent (or any direct API request) can use. They share a unified interface, are returned in OpenAI-compatible tool-call format, and are billed per usage from the same credit pool as the underlying model.

Crawl the web for up-to-date information. Returns structured results with URL citations. Available to any agent — pair it with any text model to get web-grounded answers with sources, similar to Perplexity but inside your own workflow.

Fetch and parse the content of a specific URL. Useful for summarising articles, extracting data from product pages, or feeding fresh information to agents that need to reason over current web content.

Run JavaScript code in a sandboxed Node.js runtime. Agents can write code, execute it, observe the output, and adapt — the foundation for genuinely useful data-analysis and computation agents.

Search across documents you've uploaded to a knowledge base. Powers domain-specific assistants — upload your product docs, internal handbook or research library, and any agent can retrieve relevant chunks.

Filter, sort, aggregate and transform spreadsheet data. Agents can ingest a CSV, answer analytical questions about it, and emit a transformed result. Useful for data-cleaning agents and business-analytics workflows.

Job-based PDF parsing that turns documents into structured data. Submit a PDF, get back text + extracted entities. Used by accounting, legal and HR teams to process paperwork at scale.

Vision-based structured-data extraction. Send an image of a form, receipt, or screenshot and get back a JSON object with the fields parsed. Combines well with OCR for high-throughput document pipelines.

Identify and redact personally-identifiable information from text. Useful before sending data to external models, for compliance pipelines, or for sanitising logs.

Generate natural-sounding speech from text via ElevenLabs, OpenAI TTS or Google TTS. Voice cloning supported on ElevenLabs (with consent). 30+ languages.

Transcribe audio to text via Whisper, ElevenLabs STT or Gemini STT. Multilingual, supports speaker diarisation and timestamping.

Generate images via FLUX, SDXL, DALL·E 3, Ideogram, Leonardo or Vertex. Text-to-image, image-to-image and inpainting supported. Available as an agent tool or directly via the studio.

Generate text-to-video and image-to-video via Luma Dream Machine, Runway Gen-3, Google Veo or Kling. Multi-scene sessions for longer narratives.

Current weather conditions and forecasts by location. Small but useful for travel agents, planning assistants, and any workflow where weather context matters.

Pick a model (e.g. Claude Sonnet 4.6), pick the tools the agent should be able to call (e.g. web search + code execution + file search), write a system prompt, optionally attach a knowledge base — and you have a capable assistant. Custom agents are reusable, shareable via link, and accessible through the OpenAI-compatible API.