2,289+ workflow tools across 304 enterprise apps. Managed OAuth, auto-context resolution, and server-side token injection your AI never sees credentials. Works with any MCP-compatible platform.
# Add to your MCP client config
"crawlii": {
"url": "https://mcp.crawlii.io/v1/sse"
}Post a message to a channel
{
"result": {
"items": [...],
"count": 24
}
}Agents & Workflows for Automation
API Integration for Developers
From CRM and accounting to engineering and messaging — every app your team uses, accessible to your AI via MCP.
Confluence
Wiki
Notion
Docs
Google Drive
Storage
SharePoint
Enterprise
Salesforce
CRM
HubSpot
Marketing
Pipedrive
Sales
Intercom
Support
Jira
Project Mgmt
GitHub
Version Control
GitLab
DevOps
Linear
Issue Tracking
Slack
Team Chat
Gmail
Microsoft Teams
Collaboration
Discord
Community
Xero
Accounting
QuickBooks
Bookkeeping
Stripe
Payments
Sage
ERP
Asana
Task Mgmt
Trello
Boards
Monday.com
Work OS
ClickUp
All-in-One
Raw API wrappers are provided by other platforms, leaving you to manage auth, context, and billing. The hard parts are handled automatically by Crawlii.
Your Xero tenant ID doesn't need to be known when your AI calls xero.list_invoices. Tenant IDs, workspace IDs, cloud IDs, and account scoping are automatically resolved from your OAuth connection, eliminating manual configuration.
Raw API endpoints are mirrored 1:1 by most platforms. Our tools are designed for complete workflows: names are resolved to IDs, multi-step API calls are merged into single operations, and responses are curated to deliver exactly what your AI needs.
Tokens are encrypted at rest using Fernet, automatically refreshed with distributed locking, and injected server-side. Raw tokens are never exposed to your AI. 100+ apps can be connected with one click.
Auth tokens, API keys, and context parameters are stripped from tool schemas before being sent to the LLM. Only functional parameters are made visible, ensuring zero credential exposure in compliance with OWASP MCP guidelines.
Duplicate connections are prevented when the same account is reconnected. Identity is extracted from OIDC id_tokens, /me endpoints, or email matching across 50+ providers, automatically preventing duplicate connections.
The same 300+ integrations are used to power your AI agents, app builder, enterprise search, and MCP tools. Everything is unified under one OAuth connection, one credit pool, and one dashboard. Multiple services don't need to be stitched together.
Integration tool calls deduct credits from your existing crawlii credit pool. No separate billing. No surprises.
Read, list, and get operations — fetching data costs the least.
Examples:
Create, update, send, and delete — actions that modify data.
Examples:
Search, bulk ops, and compute-heavy operations.
Examples:
Your credits cover AI agent runs, app building, media generation, and integration tool calls. No separate billing for integrations. Use more agents one month, more integrations the next — it's flexible by design.
Most platforms are just connector layers — they pipe data between apps and leave you to manage auth, context, and tool quality. Crawlii is a complete AI platform that handles the complexity for you.
Your AI never sees API keys or OAuth tokens. We handle authentication server-side with encrypted storage and automatic token refresh.
No more managing tenant IDs, workspace IDs, or cloud IDs. We automatically resolve and inject context so your AI focuses on the task, not infrastructure.
Not just API wrappers. Our tools resolve names to IDs, merge multi-step operations, and return curated responses designed for AI consumption.
Distributed Redis locks prevent race conditions during token rotation. No failed API calls, no manual retries, no downtime.
One integration powers AI agents, app builder, enterprise search, and workflows. One auth flow, one credit pool, one dashboard.
When third-party APIs change, we update our platform automatically. Your code stays the same — we handle versioning and backward compatibility.
These are real engineering problems that other platforms leave to you. crawlii handles them out of the box.
Xero, QuickBooks, and others rotate refresh tokens on every use. Concurrent tool calls can invalidate tokens. crawlii uses distributed Redis locks to ensure exactly one refresh at a time — no failed calls, no manual retries.
Your AI should never need to know a Jira cloud_id, a Xero tenant_id, or a Salesforce instance_url. crawlii resolves these from OAuth or via runtime API calls, caches them, and injects them server-side. 100+ apps supported.
Reconnecting the same Google account shouldn't create a duplicate. crawlii extracts identity from OIDC id_tokens, /me endpoints, or email matching across 50+ providers to prevent duplicate connections automatically.
"Create invoice for Acme Corp" resolves the contact name to an ID, sets sensible defaults for dates and line items, and returns a curated response — not a raw 200-field JSON blob. Each tool is designed for what LLMs actually need.
Auth tokens, API keys, and context params are stripped from tool schemas sent to the LLM (OWASP MCP01). Only functional parameters are visible. Server-side injection means zero credential exposure.
Same 300+ integrations power your AI agents, app builder, enterprise search, and MCP tools. One OAuth connection, one credit pool, one dashboard. No stitching Composio + LangChain + custom auth + separate billing.