Connectors ready to use
Agents do not work in isolation: they connect to the tools in use every day. Mail, calendar, files, messaging and code repositories can all be queried within a single conversation. Google accounts are connected through a secure authorisation, without sharing the password, and access can be revoked at any time; the other connectors are configured once for the whole company. In conversations the agent prepares sensitive actions and a person approves them. Below are the connectors already available: the ones needed are activated on request, and any other system connects through an API or an MCP server. Some tools rely on external services, indicated in the description: whether to activate them is assessed against the data perimeter chosen by the customer.

Mail and calendar
Microsoft 365 mail and calendar: the agent consults them and prepares messages and appointments, to be approved before sending.
Browsing and searching Gmail; the agent prepares drafts and sending remains the person's decision.
Checking the schedule and preparing new events in Google Calendar, directly from the conversation.
Sending email messages from the company mail server, with approval before sending.
Files and documents
Searching, opening and sharing files on Google Drive. The same documents can feed the knowledge base.
Communication
Reading and sending messages in company Slack channels, without leaving the conversation with the agent.
Project management
Searching, creating and updating issues in Jira projects, directly from the conversation.
Browsing and managing tasks, lists and folders in the team's ClickUp spaces.
Code and development
Browsing and managing GitHub repositories: files, issues and pull requests within reach of a question.
Browsing and managing GitLab repositories directly from the conversation with the agent.
Management of Heptapod repositories, the Mercurial-based variant of GitLab.
Data and integration
Connecting to any service that exposes a web API: address, method, headers and data are defined in the configuration, with protection against unauthorised calls.
Connecting to any MCP server, the open standard for AI tools: whatever the server exposes becomes immediately available to agents.
Querying any GraphQL API, to bring into the conversation the data of the systems that expose one.
A question asked in words becomes the SQL query that answers it on company data: useful for consulting a database without knowing its language. Powered by Dataherald.
Running Python and JavaScript code in a safe, isolated environment, for calculations and data processing. Powered by Riza.
Data analysis and code execution in a cloud environment kept separate from company systems. Powered by E2B.
Web search
Google searches with up-to-date results, useful for quick checks on markets, regulations and competitors.
News from thousands of sources, filtered by topic: useful for following the industry, customers and competitors. Powered by AskNews.
Clean results and concise answers, designed to be read by an agent during in-depth research. Powered by Tavily.
Search engine rankings and keyword trends, for market analysis and marketing activities. Powered by DataForSEO.
Real-time search engine results, fast and low-cost. Powered by Serper and SearchAPI.
Web search that returns page content ready to read. Powered by Jina.
Search for pages by meaning, not just by keywords. Powered by Exa.
An open-source metasearch engine that can be installed within the company perimeter, without depending on an external vendor. Powered by SearXNG.
Web search without tracking, usable with no access keys.
Results accompanied by their sources, designed for use by an agent. Powered by You.com.
Independent, privacy-conscious web search.
A search engine with its own index of the web, not based on Google or Bing. Powered by Mojeek.
AI services
Microsoft AI services: reading documents and forms, image analysis, speech-to-text and text-to-speech, clinical text analysis, for organisations already using the Azure ecosystem.
Data extraction from invoices and ID documents, object recognition in images and content moderation: useful for administration and for customer identity verification. Powered by Eden AI.
Extracting text and meaning from documents, recordings and videos, for example to make recorded meetings and training materials searchable. Powered by Nuclia.
Converting text into natural-sounding speech, in many languages, for training content or voice communications. Powered by ElevenLabs, Google Cloud or open-source models from Hugging Face.
Creating images from a description, for drafts of communication materials. Powered by OpenAI DALL·E.
Protecting agents against malicious instructions hidden in content, data leaks and unwanted content. Powered by ZenGuard.
A plain-language description of an image's content. Powered by SceneXplain.
An external knowledge base to be fed with files and web pages and then queried. Powered by CogniSwitch.
Reliable calculations, maths and scientific data, in support of technical and research functions.
Finance
Quotes, price history, financial data and news on stocks, for finance and treasury functions. Powered by Polygon.io.
Balance sheets, income statements and cash flows of listed companies, useful for analysing customers, suppliers and competitors. Powered by Financial Datasets.
The latest financial news on a stock.
Prices and trends for stocks and indices.
Research and reference
Consulting Wikipedia articles, in every language.
Search for places, addresses and businesses, useful for planning sales visits or mapping a territory. Powered by Google Places.
Technical questions and answers from Stack Overflow and the other Stack Exchange communities, in support of development teams.
How searches for a topic change over time, to gauge interest in products and markets.
Search for videos by topic, for example to follow content published about the company or its industry.
Structured information on companies, technologies and people, useful for analysing customers and suppliers. Powered by Golden.
Search for scientific papers, theses and citations, in support of research and development activities.
Search for job listings published on the web, useful for analysing the labour market and the profiles sought in the industry. Powered by Google Jobs.
Structured, verifiable data on people, places, organisations and concepts. Powered by Wikidata.
Search by image: information starting from a photo, for example to identify a product or a component.
Biomedical and scientific literature from the PubMed database, for companies in the healthcare and pharmaceutical sectors.
Search and reading of scientific preprints.
Search across scientific papers, with authors and citations. Powered by Semantic Scholar.
Search for books and bibliographic information.
Definitions and meanings of English words, in support of writing and reviewing texts in English. Powered by Merriam-Webster.
Images and data from NASA's public databases, for research and outreach activities.
Automation
Triggering IFTTT automations, to connect hundreds of services and devices.
An agent that browses the web and carries out actions in the browser, for repetitive tasks on sites without an integration. Powered by MultiOn.
Running actions from the plugins already configured for the company, such as sending messages or updating systems. Powered by Connery.
Other services
Current weather and forecasts for any location, useful for agents that plan field activities. Powered by OpenWeatherMap.
Management of applications, rules and values on the blockchain, for example in supply-chain traceability projects. Powered by AI Network.
Search for flights and availability, to organise staff business travel and related workflows, such as sending materials to a travelling salesperson. Powered by Amadeus.
Game and profile information from the Steam platform.