Fruxon logo
Fruxon

Tools & Integrations

An agent that can only talk isn't worth deploying.

Prebuilt actions on the services your team already runs on, your own REST endpoints imported from OpenAPI, direct SQL and NoSQL access, and sandboxed code execution — all callable from any step.

Integrations

Connects to the stack you already have.

Prebuilt tools for the services your team runs on — plus your own, whenever the catalog doesn't cover it.

Slack

GitHub

Jira

Google Drive

Salesforce

MongoDB

Notion

Linear

Grafana

Discord

Google Chat

Confluence

PostgreSQL

HubSpot

Airtable

Shopify

Stripe

Datadog

Sentry

GCP

BigQuery

Slack

GitHub

Jira

Google Drive

Salesforce

MongoDB

Notion

Linear

Grafana

Discord

Google Chat

Confluence

PostgreSQL

HubSpot

Airtable

Shopify

Stripe

Datadog

Sentry

GCP

BigQuery

Google Calendar

Gmail

Mixpanel

Monday

MySQL

SAP

Zendesk

Zoho CRM

Google Maps

Google Ads

Coralogix

Telegram

Apollo

Mailchimp

Calendly

Redis

Supabase

GCP Logging

gVisor

Loops

Typeform

Google Calendar

Gmail

Mixpanel

Monday

MySQL

SAP

Zendesk

Zoho CRM

Google Maps

Google Ads

Coralogix

Telegram

Apollo

Mailchimp

Calendly

Redis

Supabase

GCP Logging

gVisor

Loops

Typeform

Don't see yours? Import tools from any MCP server or request an integration.

Test against integrations without touching production

Every integration can run in Sandbox Mode. Point it at the vendor's own test environment, at the Fruxon simulator — a stateful, real-shape substitute that never leaves the platform — or at read-through, where reads hit the live API and writes are captured instead of sent. Sandbox Mode resolves per integration config, so one agent can mix sandboxed and live tools in a single run.

Your APIs and your databases, not just the catalog

Import a REST endpoint from an OpenAPI spec or configure it by hand and it becomes a tool like any other. Query PostgreSQL, MySQL, MongoDB, Redis, ClickHouse, and ArangoDB directly. Run Python in an isolated sandbox when an agent needs to compute rather than call.

Credentials live at the org, not in the agent

An integration is configured once for the organization — an OAuth grant, a connection string, a bot token — and reused across every agent. Secrets are referenced by placeholder, never pasted into a prompt. Managed OAuth flows handle the services that need them.

How it fits together

Tools are how a build becomes a system.

The same tool definitions feed skills, sub-agents, and evaluation runs. Bundle a set of tools into a skill and any agent can activate it. Route eval runs through Sandbox Mode and your golden set stops costing API credits.

Bind a tool to a specific integration config, per step

Sandbox Mode makes evaluation runs deterministic and free of side effects

Every tool call lands in the execution trace with its arguments, result, latency, and cost