A dedicated, governed runtime for your agents.
Tandem Hosted runs the governance runtime on a server that is yours alone. Channel agents for Discord, Slack, and Telegram, a knowledge base your agents can cite, and plain-English workflow authoring. You bring the AI keys.
Starter
Your own AI automation server
Announced at launch. Waitlist members get founding rates first.
- RAM: 8 GB
- vCPU: 4 vCPU
- Backups: Daily automated
- Storage: 160 GB NVMe SSD
- Tandem engine and control panel
- Knowledge base MCP (docs, Notion, URLs, files)
- Discord, Slack, and Telegram channel integrations
- Plain-English workflow authoring
- Always-on scheduled automations
- Unlimited MCP connections
- Daily automated backups
- Community Discord support
Builders
Starter plus autonomous coding and more power
Announced at launch. Waitlist members get founding rates first.
- RAM: 16 GB
- vCPU: 8 vCPU
- Backups: Daily automated
- Storage: 240 GB NVMe SSD
- Everything in Starter
- ACA autonomous coding agent
- Larger dedicated server
- Priority email support
- Extended backup retention
Pro
High-performance infrastructure and direct founder access
Announced at launch. Waitlist members get founding rates first.
- RAM: 32 GB
- vCPU: 16 vCPU
- Backups: Daily automated
- Storage: 640 GB NVMe SSD
- Everything in Builder
- Faster CPU and more memory
- Direct founder access
- First access to new runtime features
- Target response within one business day
Enterprise
Multi-user, governance, and custom infrastructure. Multi-user control, governance, SSO, audit logs, custom MCPs, and data residency options are handled as a sales conversation rather than a public plan.
Tandem is the server and the runtime. The AI that powers your agents is separate.
All tiers use your own API keys from Anthropic, OpenAI, or other providers. Your inference usage is billed directly by your AI provider. Tandem does not mark up or intermediate those costs.
That means you choose which models power your workflows, you control your inference budget directly, and you pay actual provider rates.
Dedicated runtime
A dedicated Tandem server in the launch location your tier allows, sized for your authority-layer workload instead of a shared pool.
Knowledge base for agents
Knowledgebase MCP gives agents retrieval into the sources you own — docs, Notion, URLs, and files — so answers stay grounded in your material.
Community bot, ready to go
Discord, Slack, and Telegram integrations come with Starter, and they stay included as you scale up.
Governed workflow authoring
Describe workflows in natural language and Tandem turns them into repeatable runs with visible state, approval points, and artifacts.
Built for teams that want control without the ops tax.
Solo founders and indie makers
Ship fast without inheriting a complicated self-hosting stack or a large ops burden.
Community managers and small teams
Keep your Discord, Slack, and Telegram communities answered without babysitting every question.
Engineering teams that need runtime authority
Own the box, own the keys, and keep tools, memory, approvals, and run evidence in a place your team can reason about.
Teams moving agents into real operations
Get a dedicated runtime authority layer with visible state, durable artifacts, and direct control over how agents operate.
Three steps to a dedicated Tandem runtime.
Choose your tier
Starter uses a fixed launch location. Builders and Pro open broader region coverage later.
We provision your server
The server comes online in about 10 minutes, then Tandem connects the runtime and control plane.
Connect your tools and start building
Add MCPs, ingest knowledge, and author workflows once the box is ready.
Pricing is being finalized and will be announced as each tier opens. Waitlist members see it first and get founding rates that stay with their subscription.
Starter opens first from the waitlist. Builders and Pro follow, and Enterprise is a direct conversation with the team.
BYOK across every tier
Bring Anthropic, OpenAI, or another provider and keep inference billing with the provider you choose.
Knowledge + channels included
Knowledgebase MCP and the Discord, Slack, and Telegram integrations start with Starter. They are not gated behind a higher tier.
Dedicated infrastructure
Each hosted server is provisioned on its own instance so your runtime stays isolated and predictable.
Private by default
Ownership, access, and deployment controls stay with your team instead of a shared automation provider.
Common questions, answered.
How fast can I get started?+
Join the waitlist and we onboard in order as tiers open. Once your signup is in, a dedicated server usually comes online in about 10 minutes.
What regions are available?+
Starter uses a fixed Germany (Nuremberg) launch location. Builders and Pro open with broader regional coverage as those tiers launch.
Do I need my own AI API keys?+
Yes. Every tier is BYOK. You bring Anthropic, OpenAI, or another provider key and Tandem never intermediates the inference bill.
What is included in channel integrations?+
Discord, Slack, and Telegram are included starting with Starter through the knowledgebase MCP and the hosted agent runtime.
Can I upgrade or downgrade later?+
Yes. Starter can move to Builders or Pro when those tiers open, and your active subscription keeps its billing history.
What about data privacy and hosting?+
Your server is dedicated, your keys are your own, and Tandem only runs the hosted runtime and control plane you choose.
Can I self-host instead of using the hosted service?+
Yes. Tandem Engine can still be self-hosted if you want full infrastructure control rather than the hosted service.
How does billing work?+
Hosted tiers bill monthly once they open, at the founding rates announced at launch. Your AI provider bills inference directly — Tandem only bills the hosted service.
Early customers get direct access to the Tandem team. We reply within a business day, usually faster during European business hours.
Your own Tandem server, ready in minutes.
Join the waitlist for the tier that fits. We onboard in order as tiers open, and Enterprise is a direct conversation with the Tandem team.