Local-first AI coworker
Meet TeroMate.
We believe that private AI should be affordable for every business, big or small.
Spoken by TeroMate: Hello. I am TeroMate, a local-first AI coworker. I run on the device in front of you.
Your team does not need another chatbot.
Valuable time disappears between repositories, issues, documentation, inboxes, websites and analytics tools. Traditional assistants can explain what to do. TeroMate is designed to help carry out the work - within clearly defined permissions and approval workflows.
Fragmented knowledge
Decisions, context and documentation live across repositories, chats, meetings and inboxes - and rarely in one findable place.
Repetitive operational work
Triaging issues, updating docs, preparing reports and keeping websites current consumes hours that should go into building.
Limited control over AI infrastructure
Most AI tools decide where your data lives and which model you use. For many teams, that trade-off is not acceptable.
Why we build TeroMate
Local, private AI should be affordable and accessible to every business - from the smallest studio to the largest enterprise.
No company should be left behind at the dawn of this new era. We build TeroMate so that keeping control of your data and your models is not a privilege reserved for those with the deepest pockets - it is simply how good tools should work.
One teammate. Multiple operational roles.
TeroMate is configured with role-specific skills. Some workflows are established, others are configurable, and custom integrations are shaped during the pilot.
Development & Git
Triage issues, prepare code reviews, coordinate commits and support structured development workflows.
- Issue prioritisation
- Code review preparation
- Merge request coordination
Documentation & Knowledge
Turn meetings, repositories and project information into maintainable, searchable knowledge.
- Meeting summaries
- Documentation updates
- Knowledge retrieval
Websites & Content
Keep game and company websites aligned with product updates and releases.
- Landingpage updates
- Changelogs and blogposts
- Content maintenance
Marketing & Analytics
Support teams in understanding website and campaign performance without manually collecting every data point.
- Reporting preparation
- Data analysis
- Optimisation recommendations
Social Media Updates
Draft, schedule and maintain social posts around releases, milestones and community moments.
- Post drafting
- Scheduling via Buffer
- Release & community updates
Communication
Reach TeroMate through the environments your team already uses.
- Chat
- Voice
- E-mail and collaboration tools
Operations
Track recurring tasks, prepare status updates and coordinate the next steps across tools.
- Task coordination
- Status reporting
- Approval workflows
Connects to the tools your team already uses.
TeroMate reaches into your existing stack through defined, permission-scoped connections. The integrations below already run in production inside Terovania - more are on the way.
Additional connections are scoped during the pilot. We tell you honestly what is established, what is configurable and what needs custom work - no invented integrations.
Developed inside a real agency and game studio.
TeroMate did not begin as a theoretical AI demo. Its capabilities grew from recurring workflows at Terovania Solutions and Terovania Games - including repositories, documentation, websites, analytics, communication and project coordination.
- STEP 1
A task enters the workspace
From chat, voice, e-mail or a connected tool.
- STEP 2
TeroMate gathers relevant context
Repositories, documentation, tickets and past decisions.
- STEP 3
Work is prepared or executed
Within the permissions defined for its role.
- STEP 4
A human approves critical actions
Nothing sensitive ships without sign-off.
- STEP 5
Results are documented
So knowledge stays findable for the whole team.
Your memory. Your tools. Your choice of model.
TeroMate separates where work happens from which model powers it. Choose the operating mode that fits your requirements - and change it later without starting over.
Cloud
Fastest startConnect TeroMate to existing AI subscriptions or selected model providers.
Hybrid
Most commonKeep workspace, memory and sensitive data local while routing selected tasks to cloud models.
Fully local
Maximum controlRun language models, speech processing and embeddings on infrastructure inside your organisation.
Independence, by architecture.
AI should not make your organisation dependent on one provider's pricing, availability or terms of use. TeroMate is built local-first and model-agnostic, so control over your data, your models and your infrastructure stays with you.
No vendor lock-in
TeroMate is model-agnostic. Swap between commercial APIs and self-hosted open-weights models without rebuilding workflows - so a change to one provider's pricing, availability or terms of use never becomes a change to how your team works.
Your data stays yours
Workspace, memory and sensitive project context can live entirely on infrastructure inside your organisation. No obligation to send your work to third-party clouds.
Built in Germany, for European teams
Local-first deployment lets workspace, memory and project context stay on infrastructure you choose, and keeps processing where you want it. That supports data residency and data minimisation - the parts of compliance that architecture can actually carry. The rest depends on your purposes, contracts and configuration, and we will tell you which is which.
Run it on your terms
Cloud, hybrid or fully local - the operating mode is your decision, and you can change it later. Keep speech, embeddings and inference in-house when it matters.
Keep your AI where your work is - and keep the decision about where that is yours.
Model-agnostic by design.
TeroMate runs on the model that fits your requirements - commercial cloud APIs or self-hosted open-weights models. Switch between them without rebuilding your workflows.
Further self-hosted open-weights models are being evaluated for local and hybrid deployments.
Designed to feel like part of your team.
TeroMate ships as one product: hardware and software, tuned to each other. Optional branding and physical customisation help it fit into your team and workspace - visually as well as operationally.
- Custom name and role
- Role-specific skills
- Local workspace and memory
- Microphone and speaker interface
- Optional company branding
- Logo badge and name plate
- Access through desktop app, chat and voice
The hardware
What is actually on your desk.
It arrives as one device. Everything below is part of it, and none of it is a separate purchase.
01 · Camera
Sees whether you are at your desk, scans QR codes and business cards, and reads a document you hold up to it. Recognition runs on the device - no image is sent anywhere to be understood.
02 · Microphone array
Works as a speakerphone for the room, or pairs with your headset instead. Speech becomes text on the device, so a conversation does not have to leave the building to be read back.
03 · Status display
Operating state, error messages and battery level, readable from across the desk. The three lights below it are for power and charging, the link to your inference server or cloud provider, and internet access.
04 · Compute unit
A Raspberry Pi 5 based computer in the base, with local storage and Gigabit Ethernet to your network. It runs the whole system - there is no other machine involved unless you choose one.
Shown in the configuration of a pilot device.
Sits with your team, not in place of it.
TeroMate takes the parts of a meeting nobody wants - the notes, the follow-ups, the "what did we decide" three weeks later. What it does not take is the deciding. Taste, judgement and the argument that produces a good idea stay where they belong: with the people in the room.
What TeroMate takes on
- Recording and summarising what was said
- Turning decisions into written follow-ups
- Keeping track of what is still open
- Making three-week-old discussions findable again
What stays with your team
- Deciding what is worth doing
- Judging whether the work is good
- Setting priorities and trade-offs
- The argument that produces a better idea
Every one of these can be reviewed, corrected and overruled. A summary is a draft, not a verdict.
Start with one role and one measurable workflow.
- 1
Workflow Assessment
Together we identify where recurring work costs your team the most time - and where an AI coworker fits.
- 2
Pilot Design
We define one role, one workflow and the permissions and approval points around it.
- 3
Setup & Integration
TeroMate is configured for your tools and your chosen cloud, hybrid or local AI setup.
- 4
60-Day Evaluation
We measure results against jointly defined criteria and decide how to extend from there.
The first pilot focuses on a clearly defined operational use case rather than trying to automate an entire organisation at once.
gamescom 2026 · Limited pilot programme
Apply for one of our first TeroMate pilot projects.
Following gamescom 2026, we are selecting a limited number of game studios and digital teams for a guided TeroMate pilot.
- AI Workflow Assessment
- One prioritised operational use case
- Role and skill configuration
- Integration with selected tools
- Guided implementation
- Jointly defined evaluation criteria
- Support during the pilot
- Optional cloud, hybrid or local AI setup
Pilot availability is limited and subject to a suitability assessment.
Questions teams ask before a pilot.
No. TeroMate can chat, but its purpose is operational: it works inside defined roles with concrete skills - triaging issues, maintaining documentation, preparing content or reports - always within the permissions you configure.
No. TeroMate takes over recurring, well-defined operational work so your team can spend more time on the creative and strategic parts. Critical actions always require human approval.
In its local workspace - on the device itself or on infrastructure inside your organisation. Your project context and history do not need to leave your environment.
No. You can connect existing cloud model subscriptions, run a hybrid setup, or use fully local inference. The workflows stay the same across all three modes.
Yes. Established integrations exist for common development, documentation, web and analytics tools. Additional connections are scoped as part of the pilot - we will tell you honestly what is established, what is configurable and what needs custom work.
That is configured per role. Typical approval points include publishing content, merging code, sending external communication and any change to production systems. Read-and-prepare tasks can run more autonomously.
Yes. Several devices can share a single local inference server, so teams can add coworkers without duplicating hardware for model execution.
Yes - name, role, logo badge and name plate can be tailored so TeroMate visually belongs to your team.
We run a workflow assessment, design one focused use case, set up the integration and evaluate results together over roughly 60 days against criteria we define jointly at the start.
No. Game studios and publishers are where TeroMate grew up, but agencies, software companies and other digital teams share the same recurring workflows it is built for.
What happens next
It starts with one workflow, not with a rollout. We look at where recurring work actually costs your team time, and say honestly whether an AI coworker belongs there. Sometimes the answer is no - and that is a much cheaper thing to find out in a conversation than in a pilot.
Which recurring task should your first AI coworker take off your team's desk?
Start with one workflow.
Tell us which recurring task costs your team the most time. We will come back to you with an honest assessment of whether TeroMate fits it - and say so if it does not.