> ## Documentation Index
> Fetch the complete documentation index at: https://docs.2501.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# FAQ

> Quick answers to common questions about 2501

## General

> [accordion] What does 2501 actually do?
>
> 2501 runs autonomous AI agents on your infrastructure. You point it at machines, give it a few rules about how your environment works, and it handles tickets from your ITSM - investigating, fixing, and reporting back in the ticket itself.

> [accordion] Does it work without a ServiceNow integration?
>
> Yes. ServiceNow is the default gateway and what most deployments run, but custom-built gateways are available for other ticketing systems - contact your account team. You can also drive 2501 from the Command Center UI directly.

> [accordion] Is there a free tier?
>
> Yes - a built-in free tier caps your tenant at **50 lifetime tasks** and **10 hosts** (gateways are unlimited). Once you apply a license, those caps are replaced by your contract's. See [Licensing](https://docs.2501.ai/0.19/configure/licensing.md).

## Agents and safety

> [accordion] How do I keep agents from running destructive commands?
>
> Three layers, used together:
>
> 1. **[Permissions](https://docs.2501.ai/0.19/risk/permissions.md)** - give agents the least-privileged credentials that still let them work.
> 2. **[Blacklists](https://docs.2501.ai/0.19/risk/blacklists.md)** - programmatic blocks for specific command patterns, always on.
> 3. **[Read-Only Agents](https://docs.2501.ai/0.19/risk/read-only-agents.md)** - set a specialty to investigate-only for read-only operation by default. An explicit `@2501:remediate` ticket tag overrides that setting for the job.

> [accordion] What's the difference between investigate and remediate?
>
> Investigate is read-only - the agent diagnoses and reports without changing anything. Remediate lets the agent act. Without a tag, a ticket's tasks remediate unless their specialty is investigate-only. Set per ticket with `@2501:investigate` or `@2501:remediate`, or set a whole specialty to investigate-only. See [Read-Only Agents](https://docs.2501.ai/0.19/risk/read-only-agents.md).

> [accordion] What's the difference between a Specialty and an Operational Rule?
>
> A **Specialty** shapes how the agent *thinks* about a domain - vocabulary, priorities, approach. An **Operational Rule** tells it what is *true* in your environment and what that implies. Specialty: "Kubernetes Operator". Rule: "On this cluster, rollbacks go through `helm rollback`, never re-applying manifests."
>
> Rules outrank specialties in conflicts. See [Prompting an Operational Rule](https://docs.2501.ai/0.19/prompting/operational-rule.md).

> [accordion] Why is there a 20-task limit per job?
>
> The results of every task go into the planning of the next turn and into the final answer, so a job's prompts grow with each task. 20 tasks keep them within the model's context, and a patch on 20 hosts still fits in one turn. If a job reaches the limit while the ticket still needs work on more hosts, the job ends as partial. See [Agentic Flow](https://docs.2501.ai/0.19/understand/agentic-flow.md#the-20-task-ceiling).

## Models and providers

> [accordion] What LLM providers does 2501 support?
>
> Native: OpenAI, Anthropic, Mistral, DeepSeek, Cohere, OpenRouter, TogetherAI. Compatible: any OpenAI- or Anthropic-shaped API via `openai-compatible` / `anthropic-compatible`, and Azure OpenAI (including APIM). Self-hosted endpoints (vLLM, Ollama) work through the compatible types. See [Providers](https://docs.2501.ai/0.19/configure/providers.md).

> [accordion] Can I use different models for different agents?
>
> Yes. Each agent picks its **main** and **secondary** engine from the tenant catalog. The tenant has its own defaults for gateway routing and the AI Assistant. See [Engine & Agents](https://docs.2501.ai/0.19/understand/engine-agents.md).

> [accordion] Why two engines per agent?
>
> The main engine does the work (commands, file edits). The secondary engine plans, watches, and judges every command for compliance and read-only enforcement. Separating execution from oversight improves both accuracy and safety. See [Engine & Agents](https://docs.2501.ai/0.19/understand/engine-agents.md#the-two-engine-agent).

## Deployment

> [accordion] Docker Swarm or Kubernetes?
>
> Either. The same signed chart installs on Kubernetes (any distribution: vanilla, OpenShift, k3s, EKS, etc.), applied with Helm, ArgoCD or `kubectl`, and on Docker Swarm, rendered into one `compose.yaml` for `docker stack deploy`. See [Overview](https://docs.2501.ai/0.19/deployment/overview.md).

> [accordion] How do I upgrade 2501?
>
> Apply the chart of the newer build with the same `values.yaml`, after reading the Upgrade actions of every release you cross and reviewing the diff. See [Upgrade](https://docs.2501.ai/0.19/deployment/upgrade.md).

> [accordion] Where do I configure secrets?
>
> The installation's own keys, its database connection and the first LLM API key are secrets you create in your cluster or Swarm; `values.yaml` only names them. Further LLM providers and their keys are configured in **Command Center**. Per-host credentials (SSH, WinRM, gMSA, vault paths, PAM accounts) live in **Command Center → Credentials**, encrypted at rest. Agents only access them programmatically - values are never displayed in the UI after creation. See [Credentials](https://docs.2501.ai/0.19/configure/credentials.md).

## Knowledge and rules

> [accordion] What file types can I upload as knowledge?
>
> PDF, DOCX, CSV, and Markdown. Max 200 MB per file. PDFs with diagrams need a multimodal model on the tenant to capture the visual content. See [Knowledge](https://docs.2501.ai/0.19/configure/knowledge.md).

> [accordion] What if my uploaded knowledge contradicts a rule I wrote by hand?
>
> Manual rules always win. The auto-extracted rule is replaced when you re-upload the same document, but your hand-written rules are untouched. See [Working with Knowledge](https://docs.2501.ai/0.19/best-practices/knowledge.md).

> [accordion] How do I know which rules reached a task?
>
> Each task's detail page in Command Center shows a **matching trace** - which operational rules were applied, which were skipped, and why each skip happened. Use it to debug rules that are too broad or too narrow.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.