What each screen is for, what to do there, and what outcome to expect.

Home

The dashboard you land on. Glance here to see if everything's running, then jump to wherever you need to go.

Operational front door for instance health, routing status, host pressure, active work, and high-value jumps into the rest of the app.

• Check whether any instances are ready for requests.
• Look for machine pressure before starting more models.
• Jump to Runtime, API Access, Chat, or diagnostics when something is not ready.

Models

Find, download, inspect, and approve the model files that Ai Keeper can run.

The model inventory and supply-chain review surface.

• Library: inspect installed assets, formats, sizes, capabilities, runtime fit, and dependent instances.
• Downloads: search Hugging Face, review model cards, and download MLX or GGUF weights.
• Trust: review static supply-chain findings and create explicit, inventory-bound launch overrides.

Runtime

The engine room. This is where downloaded models actually start running so you can talk to them.

Where model assets become running API-capable instances.

• Instances: create, start, stop, edit, diagnose, and tune runtime instances.
• Inspect availability, routing, diagnostics, and resource pressure for each running server.

Chat

The main place to talk to the AI. Type, attach files, get answers — like ChatGPT but on your machine.

Primary conversation workspace with streaming, attachments, tool calls, approvals, session controls, slash commands, compaction, exports, and model recommendations.

• Use it for normal work, prompt testing, file analysis, multimodal inputs, and guided tool use.
• Use /commands to see every built-in and custom slash command.
• Use /compact or auto-compaction for long sessions.

Media

Create and improve pictures, video, and music on your own Mac.

Media Studio combines local image, video, music, enhancement, gallery, and dependency-setup workflows in a managed Python environment.

• Image, Video, Music: generation controls tailored to each output type.
• Enhance: upscale, remove backgrounds, and perform supported edits.
• Memory Guard and Video Guard estimate the cost of a run before it starts and cap the frame count that fits your machine.
• Gallery: review outputs and metadata. Setup: install or repair the local media runtime.

Engines

The installer. Every serving engine and supporting tool the app needs lives here, with a button to install, update, or repair each one.

Dependency and engine lifecycle surface. Shows required vs optional dependencies, resolved paths, install source, version policy, and release channel per engine.

• Fix Required installs everything marked required in one pass — the first thing to run on a new install.
• Install Optional… lets you pick extra engines and tooling from a checklist.
• Update All, plus per-row Install, Update, Repair, Uninstall, and Re-check.
• Release channel pins stable vs pre-release builds per engine; Automatic engine updates keeps managed engines current on its own.
• Output Log shows exactly what the installer ran when something fails.

Activity

The evidence center for traffic, failures, health, history, and machine pressure.

Requests, Flight Recorder, Logs, Health, Usage, Sessions, Exec Sessions, and Monitor are collected here. Flight Recorder follows metadata-only request and egress paths without recording prompts or credentials.

Settings

Configure how this Mac connects, where data lives, which providers and tools are available, and how much autonomy workflows receive.

Eight app destinations plus a searchable operator catalog.

• General, Connection, Providers, Coding Harness Control Plane, Storage, Language, Work Modes, API Access.
• Operator catalog: security, connectivity, device, and operator tools in the same searchable Settings sidebar.

Knowledge

Everything the AI can "know" about you and your work — facts it remembers, files it can search, notes you've written, even people in your contacts.

The context layer chats and agents can draw from.

• AI Context: active system prompts and shared prompt shaping.
• Skills: installed and draft skill bundles.
• Memory: persistent facts, manual entries, auto-extraction, markdown export, delete controls.
• Audit: which memory facts were injected into recent prompts, where, and why.
• Documents: RAG files, folder indexing, chunking, search, selected embedding instance.
• Wiki: curated pages, time-aware claims, current-truth pointers, freshness lint, revisions, links, backlinks, and tags.
• Ingestion: connector-inbox sources reconciled through reviewed create/replace/no-op, enrichment, contradiction, synthesis, and retirement proposals.
• Dreaming: idle-generated diary entries, memory-palace pages, and insights.
• Contacts: people, groups, tags, and reachability context.
• Obsidian: vault connection, note search/import, filesystem watching with a periodic fallback, source retirement, and optional conflict-checked write-back.

Agents

A team of AI agents working together on a job. You set up the team once and run it whenever you need.

Repeatable multi-agent work.

• My Agents: create and edit reusable agent definitions.
• Workspaces: group agents, launch work, monitor messages, resolve confirmations, and review outcomes.
• Agents can have roles, models, ports, skills, memory scopes, tool access modes, standing orders, hooks, and prompt previews.

Automation

Schedule the AI to run jobs by itself — on a timer, when something happens, or on a regular heartbeat.

Converts reliable prompts, agents, and tools into scheduled or event-driven work.

• Flows: tasks, hooks, orders, advanced jobs, heartbeat, background jobs, workflows, and cron editor.
• Triggers: inbound webhook endpoints.
• Use approvals for any automation that can write files, run commands, use browsers, call channels, or affect external systems.

Extensions

Add-ons. Connect the AI to outside services like Slack or Telegram, manage installed plug-ins, or hook in extra tools.

Installed capabilities and external surfaces.

• Installed: package runtime health, setup, update, repair, and uninstall.
• Channels: shared inbox, Telegram, WebChat, auto-reply, DM Access, broadcast, accounts, and locations.
• Plugins: app plugin lifecycle and contributed capabilities.
• MCP: server configuration and discovered tools/resources.
• Skills and ClawHub: manage reusable behavior bundles and browse installable packages.

Lab

A sandbox to try prompts and compare models before committing them to a real workflow.

Where you test before depending on a setup.

• Playground: raw API request testing.
• Responses: the stateful Responses API, compacted history payloads, and WebSocket transport against local runtimes.
• Compare: side-by-side prompt evaluation.
• Benchmark: TTFT, throughput, and latency checks.
• QA Lab: test suites and regression probes.
• Transcription: timestamp local audio, review speakers, and promote material into Knowledge.
• Harness: inspect experiments, causal traces, context budgets, tool contracts, learned rules, and scratch workspaces.

Chat session tools: Browser and Canvas open from Chat so web research and rendered artifacts remain attached to the active conversation.