Connect Splynx to your AI tools

Use the Splynx MCP to manage and analyze your network and customers, or build needed integrations with your existing AI setup. Compatible with Claude, ChatGPT, Gemini, and any MCP-compatible LLM.

Try it free

WISPA 2025 Award Winner Service of the Year for ISPs

What is an MCP?

MCP (Model Context Protocol) is an open standard that lets AI assistants connect to external tools and data through one common interface, instead of a custom integration for every service. Splynx uses MCP in two ways.

MCP SERVER

Claude

ChatGPT

Cursor

Gemini

Perplexity

+ Any MCP client

Two MCP servers, two different jobs

One connects an assistant to your Splynx instance. The other teaches an AI coding assistant how Splynx add-ons are built.

Splynx MCP Server · For your team

Run Splynx by asking your AI assistant

Good to know

Built into Splynx 6.0 - one secure endpoint on your own instance

Works with Claude, Cursor, ChatGPT and any MCP-compatible assistant

Uses the permissions you already manage - every AI action is logged

How it works

Describe what you need in plain language, in any language

The assistant finds the right operation, runs it, and reports back

Read the setup guide 5-minute setup · no coding required

Claude - connected to Splynx

Show the unpaid invoices for customer John Smith

searched finance - ran list_invoices

Found John Smith (#1042) - 2 unpaid invoices:

INV-2481 - Jul 2026$49.00

INV-2532 - Aug 2026$49.00

"Who's overdue this month?" "Sessions for customer #88" "Leads this week, by source" "Change tariff to Fiber 300" "Block internet for customer #42" "Everyone on Fiber 100 in {location}" "Create a ticket for #1042"

Use it for

Building AI agents Workflows & automation Reports & insights Search & data lookups Automating key business tasks

You get

Routine work done by asking, in any language Bound to existing roles & API keys - nothing more Confirmation before writes and deletes Every AI action audited and flagged

claude code - splynx-addons

› Scaffold an add-on "SMS Notifications" that reacts to customer and invoice events

⏺ scaffold_addon - generate_hook_controller

sms-notifications/

├ controllers/InstallController.php ✓

├ controllers/HookController.php ✓

├ config.json ✓

└ config/web.php ✓

Hooks wired for customer/create and invoice/create.

doc topics

code generators

guided workflows

Add-ons MCP Server · For developers

Build Splynx add-ons with AI that knows the architecture

Good to know

A developer tool that runs on your own machine or infrastructure

Loads the Splynx add-on architecture, hooks and API reference into your AI coding assistant

Never connects to your live instance or customer data

How it works

Point Claude Code, or any MCP-compatible coding tool, at the server

Describe the add-on or integration you need

It scaffolds the files and wires the events the way Splynx does - you review and deploy

Read the developer guide Runs locally · open source on GitHub

Use it for

Building custom integrations & add-ons New Payment gateway and network integrations Accounting & ERP sync E-invoicing & tax compliance Any other integration or sync

You get

Working skeletons in one request, not days Code that matches Splynx's own Yii2 patterns Guidance through the parts that are easy to get wrong Zero access to your live instance or customer data

FAQ

What is Splynx MCP?

MCP (Model Context Protocol) is an open standard that lets AI assistants like Gemini, Claude, ChatGPT, or Cursor connect securely to external systems and use them as tools. Splynx offers MCP through two separate servers, each built for a different job.

The Splynx MCP Server connects an assistant to your live, running Splynx instance, so it can look up customers, check invoices, review tariffs, and pull reports using the same permissions as the account it signs in with. The Splynx Add-ons MCP Server is built for developers instead: it gives a coding assistant the add-on architecture, API reference, and code generators needed to build and extend Splynx, without ever touching a live instance or customer data. Together, they let you use Splynx by asking, and build for Splynx with an assistant that already knows how it's put together.

How is my data protected?

Like any MCP server, ours gives a connected AI assistant access to data through the API, scoped to whatever permissions you grant it. For the live-instance connector, that means it can read and act on real customer data, but only within the permissions of the account or API key it's connected with, and every action it takes is logged for review. The add-ons dev tool works differently: it never touches live data at all, only your own codebase and Splynx's bundled documentation.

Because any MCP connection is still a channel into your business data, we recommend connecting only enterprise-grade AI platforms whose license terms confirm your conversations aren't used for model training and aren't shared with third parties.

Which AI clients does it support?

Both servers speak the standard Model Context Protocol, so in principle any MCP-compatible AI client can connect to either one. In practice, the live-instance connector has been tested with Claude Desktop, Claude Code, and Cursor, while the add-ons dev tool is built and demonstrated specifically around Claude Code, since that's the primary assistant developers use for scaffolding Splynx add-ons today.

Is this available to my customers?

No. Both servers are internal tools meant for people who work at your ISP, not for the customers you serve. The live-instance connector is intended for admin staff who already have a Splynx account or API key, and the add-ons dev tool is for developers building software for Splynx. Neither is designed, secured, or licensed as a customer-facing support channel.

Can the AI make changes without my approval?

For the live-instance connector, operations that modify or delete data are flagged, so the assistant confirms with you before carrying them out. If you want an extra layer of safety, you can also connect with a read-only API key, which removes the ability to write at all. The add-ons dev tool works differently: it only writes to your own project files and never touches a live Splynx instance, so as with any AI-generated code, we recommend reviewing what it produces before you deploy it.

Do I need a live Splynx instance to use these?

That depends on which server you're using. The live-instance connector needs one, since that's exactly what it connects to. The add-ons dev tool doesn't; it's self-contained and runs on the Splynx add-on architecture and documentation rather than a live deployment. You'd only need an actual Splynx instance later, to install and test whatever add-on you end up building.

Join 1000+ ISPs and build your tech stack on Splynx today

Try it free