Support
Frequently Asked Questions
Answers to the most common questions about DigiSwarm. Can't find your answer? Get in touch and we'll get it for you.
Code & Workflow
Your Team & Control
AI Models
Code & Workflow
How do agents access my code?
Each agent runs in its own isolated container with a scoped GitHub token and read/write access to your repository. Agents read your codebase, write code on their own branches, and open pull requests — all within the permissions you grant. Nothing runs on shared infrastructure.
How is my code kept secure / isolated?
Every customer gets a dedicated, isolated Kubernetes namespace — your agents, secrets, and data are fully separated from other tenants, with no shared compute. Agents use scoped GitHub tokens and run with least-privilege access.
Do agents work inside my existing GitHub workflow?
Yes. Agents work directly inside GitHub — the same issues, pull requests, and labels your team already uses. There's no new tooling to adopt and no workflow changes required.
Do I need to manage infrastructure?
No. DigiSwarm runs the platform — the Kubernetes namespaces, agent containers, orchestration, and scaling are all handled for you. You connect your GitHub repository and your agents go to work. For Enterprise, you can also run DigiSwarm on your own infrastructure (see Enterprise).
Your Team & Control
What does Paul do?
Paul is your AI chief of staff. He turns your direction into action — writing and assigning issues, dispatching the right agents to each task, reviewing pull requests, and reporting back. You brief Paul; Paul runs the team.
How many agents can run at once?
It depends on your plan. Starter runs one agent at a time; Growth runs up to 5 in parallel; Scale runs up to 25; Enterprise is custom. Agents are staffed on demand — when there's no work, no agents run. (See pricing.)
Can agents work autonomously, or do I stay in control?
That's your call. DigiSwarm runs anywhere from fully supervised to fully autonomous — you set the level. Keep a human review gate on every merge, let the swarm run on its own, or land anywhere in between. Both ends are first-class: full oversight and full autonomy are equally supported.
What does “staffed on demand” mean?
Agents are spun up only when there's work to do. When a task lands, the right agent starts; when there's no work, no agents run — so you're never paying for idle agents or idle token spend.
What counts as an “active agent”?
An active agent is one currently working on a task. Your plan's agent limit is about how many can run at the same time — not a cap on total work. Agents are staffed on demand, so they only count as active while they're actually working. Paul, your chief of staff, doesn't count against your agent limit.
AI Models
Which AI model do agents use?
The model of your choice. DigiSwarm supports Claude, OpenAI/GPT, DeepSeek, and self-hosted Qwen3 — you choose which model powers your agents.
Can I use more than one model?
Yes. DigiSwarm is multi-model — you can run different agents on different models, and switch models without changing your agent setup, tooling, or skills.
Plans & Enterprise
Can I change plans later?
Yes. You can upgrade or downgrade at any time as your needs change — scale up during a heavy sprint, scale back between releases. See plans and pricing.
Can I run DigiSwarm entirely on my own infrastructure?
Yes. Enterprise customers can run DigiSwarm fully air-gapped on their own servers or VPC, with a self-hosted DeepSeek model. Your code, agents, and model all stay inside your walls — nothing is sent to an outside API, and there are no per-token costs.
Is there an audit trail?
Yes. Every agent action, configuration change, and API call is logged and exportable — a full record of what your agents did and when, for your security and compliance teams.
What about RBAC / access control?
Role-based access control lets you define who can configure agents, view dashboards, approve merges, and manage billing — mapped to your existing org structure. Provisioned as part of Enterprise deployments.
Still have questions?
Tell us what you're building and we'll get you the answers you need.