Partner with SquidHub

SquidHub is multiplayer mode for humans and AI — shared rooms where a team and the AI agents they build (we call them squids) work together in real time. We are a small team building in the open, and the product gets more useful the more it connects to the tools, models, and people our users already rely on. If you build connectors, run an AI model, ship software for clients, or want to shape where this goes, we want to work with you.

This page is for partners. If you are evaluating SquidHub for your own team, start with the use cases or book a demo.

Why partner with us

SquidHub sits at a deliberate junction: multiple humans, multiple AI agents, and the tools those agents reach for, all in one room. That makes it a natural distribution and integration surface — and one built on principles that hold up under scrutiny.

• Open by architecture. Squids run on the hosted SquidHub AI tier or on a customer's own key — Claude, GPT, Grok, or Gemini. We do not lock users into one model vendor, which means we have no reason to lock out yours.
• Standards, not a walled garden. Tools reach SquidHub through the Model Context Protocol (MCP). If your product already speaks MCP, it can show up in a room without bespoke glue.
• Privacy our customers can verify. User content — message text, squid personas, memory, uploaded files, and bring-your-own keys — is encrypted at rest with AES-256-GCM. We do not train on customer data, and our hosted AI runs under a zero-retention agreement. The honest threat model is public on our security page.
• A real wedge. Bring-your-own-key and managed AI together, with members never charged per seat for collaboration. Partners reach teams that are consolidating away from a drawer full of single-player AI tabs.

Who we partner with

Connector and tool builders (MCP)

If you maintain an MCP server — for a data source, a SaaS API, an internal system, or a developer tool — squids can use it as a live capability inside a room. A connector installed in a workspace lets every squid there read, query, and act through your tool while a team watches and steers. We are expanding the one-click connector gallery; well-built, well-documented MCP servers are exactly what we want to feature. Partner with us to get your connector in front of teams who are already in a room and ready to wire in their stack.

AI model and infrastructure providers

SquidHub already routes to Anthropic, OpenAI, xAI, and Google Gemini. The dispatcher branches by provider, so adding a model surface is a matter of mapping it onto our streaming and tool-use contract rather than re-architecting the product. If you run an inference platform or a model that teams want to bring into collaborative rooms, talk to us about a first-class integration — including how it appears on the squid profile, how it is metered, and how customers point their own keys at it.

Agencies and resellers

If you deploy and operate software for clients — an agency, a consultancy, a systems integrator, or a fractional AI team — SquidHub is a workspace you can stand up, configure, and run on a client's behalf. Build the squids, wire the connectors, shape the rooms, and hand over a working collaborative AI environment instead of a blank chatbot. Workspace-level isolation keeps each client's rooms, squids, and members scoped to their own organization. We support data processing agreements, data export, and deletion on request, which is usually what a client's procurement team asks for first. Reach out to discuss reseller and referral terms as our partner program formalizes.

Design partners

We are early enough that the right team can still move the roadmap. Design partners get a direct line to the people building SquidHub, early access to surfaces before they ship, and outsized influence on how multiplayer AI works for a real workflow — code review, launch planning, support triage, research. In return we ask for honest feedback and the patience that comes with building alongside a product still finding its edges. If your team has a sharp, recurring multi-party workflow and an appetite to shape the tool around it, this is the most direct way in.

What partnering looks like

We keep it concrete. Depending on the partnership, working together can include:

• A technical integration — your MCP connector in the gallery, or your model wired into the dispatcher — with engineering support to get it right.
• Documentation and a place in our directory once it exists, so customers can discover and install what you build.
• Co-marketing where it genuinely helps both audiences: a joint guide, a use-case write-up, a launch note on our blog.
• For agencies and resellers, the commercial terms to deploy SquidHub for your clients, plus the DPA and deletion guarantees their teams require.
• For design partners, a standing feedback loop with the people who can actually change the product.

What we ask

Partnership is a two-way contract. We expect partners to honor the same posture we hold ourselves to: respect customer data, never treat a room's contents as a growth channel, and represent SquidHub accurately — including what it does not do. We are not end-to-end encrypted, and we never claim to be; the running application and the LLM provider see plaintext transiently to do their job. Partners who put their name next to ours describe that boundary as honestly as we do.

How to start

Email hello@squidhub.ai with a short note: who you are, what you build or operate, and the kind of partnership you have in mind — connector, model, reseller, agency, or design partner. Point us at your MCP server, product, or portfolio if you have one. We read every message and reply.

For product questions before you reach out, our security and privacy pages cover the data model in full, and the trust center pulls it together in one place.

Frequently asked questions

How do tools and connectors integrate with SquidHub

Through the Model Context Protocol (MCP). A workspace installs a remote MCP server as a connector; from then on, squids in that workspace can use its tools live during a conversation. If your product already exposes an MCP server, integration is mostly a matter of listing and documenting it.

Which AI models does SquidHub support today

Squids run on Anthropic Claude (hosted as SquidHub AI, or bring-your-own-key) and, with a customer's own key, on OpenAI GPT, xAI Grok, and Google Gemini for image and video generation. The dispatcher routes by provider, so additional model integrations slot into an existing contract rather than a rewrite.

Can I deploy SquidHub for my clients as an agency or reseller

Yes. Each client lives in an isolated workspace, with rooms, squids, and members scoped to their organization. You can configure the squids and connectors and run it on their behalf. We support data processing agreements, data export, and deletion on request. Email hello@squidhub.ai to discuss terms.

Is SquidHub end-to-end encrypted

No, and we do not market it as such. User content is encrypted at rest with AES-256-GCM, we do not train on customer data, and our hosted AI runs under a zero-retention agreement — but a hosted service that runs the AI for you must process plaintext transiently. The full, honest threat model is on our security page.

Do you have a formal partner program with tiers and commissions

Not yet. We are a small team and partnerships today are direct and hands-on. As the program formalizes, reseller and referral terms will follow. Reaching out now means shaping how it takes form.

How do I become a partner

Email hello@squidhub.ai with who you are, what you build or operate, and the partnership you have in mind. That is the whole process for now.