Skip to content

AI Copilot + Mobula

The AI Copilot is an assistant that lives in a global dock on the right of every screen. It answers questions about your cluster by reading Mobula — KubeManta's live model of what exists in your cluster, how it is connected, and what is currently wrong with it. Because answers come from Mobula rather than from a generic prompt, they cite the actual resources they were derived from, and you can click a citation to jump straight to that resource.

Tiering

The structural Mobula surface (/graph/*) is Free and works with AI disabled — it is deterministic Kubernetes/DB topology, the same data behind the free Cluster and Network panels. The Copilot chat is Pro (ai feature): Free installs get a small daily teaser quota; a licensed ai cluster is unlimited.

Mobula

Mobula is assembled from the signals KubeManta already collects — workload topology, failure cascades, network paths, resource relationships, health signals and container-scanning results — together with KubeManta's own operational memory, such as the audit trail, past diagnoses and alert history.

The result is one connected picture of the cluster rather than a pile of separate views: an alert, the pod it fired on, the workload that owns it, the node it landed on and the finding against its image are all reachable from one another.

Two properties matter operationally:

  • It is built on demand. There is no graph database to run, scale or back up, and nothing to keep in sync — Mobula is assembled per request from live cluster state.
  • It respects your boundaries. Namespace allowlists, system-namespace hiding and secret redaction apply to Mobula exactly as they do everywhere else in the product.

The same model powers the Mobula cluster map in the UI and grounds the Copilot's answers.

Mobula API (Free)

Endpoint Auth Returns
GET /graph/subgraph?focus=&hops=&budget= Authenticated A token-budgeted subgraph around a focus node.
GET /graph/entity/{id} Authenticated Node detail + a deep-link to the relevant UI view/namespace.

These are _require_auth only — not gated on AI or a license. They exclude AI-derived memory nodes and respect the namespace allowlist, system-namespace hiding, and _scrub redaction.

How the Copilot retrieves context (Pro)

When you chat with the Copilot it navigates Mobula — following the connections around whatever you asked about — instead of dumping the whole cluster into the prompt. In practice that means a question about one failing pod pulls in that pod's events, its owner, its recent findings and little else.

Retrieval is bounded by design: it walks a limited distance from the starting point and works to a token budget, so a question about one workload cannot quietly turn into a cluster-wide scrape. Every step re-checks namespace scope against the AI Guardrails — an out-of-scope target returns an error, never raw data — and any log content is redacted before the model sees it.

The same read-only retrieval is available to external agents through the MCP server.

AI Context — your docs, retrieved on demand

The AI Context tab (sidebar, AI group) is an operator-curated markdown knowledge base — org runbooks, conventions, escalation paths — that the AI retrieves on demand via a search_context tool instead of re-deriving them every conversation. Token-cheap keyword retrieval: the AI searches, then pulls the full text of only the doc it needs. CRUD is admin-gated (GET/POST/PUT/DELETE /context/docs); docs are capped in size and count so the knowledge base stays prompt-friendly.

Save a conversation to AI Context

The Copilot's Save to context button (POST /context/docs/save-conversation, admin) distills a chat thread into a reusable context doc. One guardrail-enforced AI pass writes a compact summary — and only that summary is indexed for search_context retrieval. The full verbatim transcript is preserved for audit in a collapsed block below a <!-- KM_TRANSCRIPT_BELOW --> sentinel, but is deliberately excluded from search scoring so it never inflates results or snippets. The net effect: a future question retrieves the distilled outcome instead of replaying the whole history (roughly ~230× fewer tokens on a sample thread). If no AI key is available, a deterministic extractive summary is saved instead. The same AI Guardrails (data residency, cost caps, redaction) apply to the summarization pass.

Include a workspace file in a chat

From the Log Workspace, the Include in AI action attaches the open file to the Copilot dock as an explicit attachment (it goes to the AI dock, not the PTY terminal) and persists with that conversation — so you can hand the AI a log dump, a manifest, or captured output as first-class context without pasting it into the prompt.

Using the dock

  • ⌘K / Ctrl-K opens the Copilot from anywhere.
  • One session follows your tabs — ask about whatever you're looking at without re-establishing context.
  • Answers use [[node-id]] citations; click one to deep-link to that resource's view.
  • Inline Explain / Fix triggers appear on individual surfaces (a failing pod, a finding) to seed a scoped question.

Write safety

The Copilot reuses the existing chat loop, so it inherits every AI guardrail with zero new privileged surface:

  • The master AI kill-switch, data-residency allowlist, and cost caps all apply unchanged.
  • Any write the AI proposes still flows through the approve-before-apply token gate — nothing mutates the cluster without an explicit human approval, blast-radius check, and circuit-breaker check.

See AI Guardrails for the full policy surface.

Free quota vs Pro

Free Pro (licensed ai)
Structural graph (/graph/*) ✅ ✅
Copilot chat Daily teaser quota (default 15 prompts/actor/day) Unlimited (never counted)

When a Free actor exceeds the daily quota, the API returns a structured 402 upsell that the dock renders as a Pro card — not a raw error. The quota is tracked per actor per day.

Prompt audit

When Prompt Audit is enabled in AI Guardrails, Copilot chats (both Anthropic and OpenAI paths) write the scrubbed prompt + response to the compliance log — the Copilot is a first-class audited AI surface.