Choosing Your LLM & Coding Harness

Every agent needs an AI to think with. Agentum comes with one already chosen and already working, covered by the usage in your plan — this page explains what that is, how we pick it, and how to swap in your own key instead.

Agentum makes two independent choices here, and it’s worth keeping them straight:

  • The LLM provider is the model your agents talk with — it powers chat, summaries, digests, and the intelligence inside every agent.
  • The coding assistant is the tool that writes agent code when you use Create, Improve, or Sidekick.

They’re configured side by side on the LLM Provider card in Settings → Connections, and they don’t have to match: you can chat through Gemini and build with Codex, or run everything on the built-in default.

What Agentum AI actually is

Agentum AI isn’t a model we built. There are a handful of companies in the world who train these things properly, and we are not one of them — pretending otherwise would be a strange thing to sell you.

What we do instead is make the choice practical. Agentum AI gives you a curated set of models with a sensible default, already wired in and paid for by your monthly usage. There is no provider account or API key to set up, and you can change the model from the same panel if you want to.

Routine work does not need the most expensive model. Agentum sends lightweight background summaries to a lower-cost tier and keeps stronger reasoning for the paths that need it. Your selected model remains the primary choice for chat and Agentum Code.

You see usage, not token arithmetic. Agentum handles the provider rates and metering behind one balance. The app still shows the provider and exact model, so the convenience does not hide what is running.

What a job costs can change. These services reprice from time to time, and when that moves what something costs in usage, we tell you before it happens rather than after.

Used up your month? Reset it. You get alerts at 80% and 95%. At 100% your agents pause. A reset adds a full month’s usage to your balance right away, on top of whatever you have left. Pricing has the detail.

If you want to know what’s running right now, the app tells you. Settings → Connections shows the live provider and model in the corner of the LLM Provider card.

Where to find it

Everything above lives in one panel. Open Agentum, go to Settings, then the Connections tab.

The badge in the corner always names the provider and the exact model version currently in use — so "which AI am I on?" is never a guess.

The LLM provider

Your options

Provider What you need Notes
Agentum AI (default) Nothing No key, no signup — your plan’s monthly usage covers it. Agentum curates the models, sets the default, and pays the provider.
Claude (Anthropic) Anthropic API key (sk-ant-…) Get one at console.anthropic.com.
Gemini (Google) Gemini API key (AIzaSy…) Get one at aistudio.google.com. Also enables voice transcription.
DeepSeek DeepSeek API key (sk-…) Get one at platform.deepseek.com.
MiniMax API key and GroupId From minimaxi.com.
None — Disables LLM features. Agents that don’t use AI keep working.

Agentum AI is the zero-setup path. There’s nothing to paste: your license itself is the credential, and Agentum’s servers meter and forward requests. If you already pay for an API key elsewhere — or want a specific model — bring your own key instead.

It’s included, not unlimited. Agentum pays the model bill, and your plan’s monthly usage allowance covers everything Agentum funds — chatting with Agentum, onboarding, building agents, the thinking inside the agents you run, and Agentum Code sessions. The app’s Account card always shows where you stand. Alerts land at 80% and 95%; at 100% your agents pause rather than run up a bill, and you can reset the month, move up a plan, or connect your own key on this same screen. There is no overage charge, ever.

Picking a model

Each bring-your-own provider has a model field with suggested picks as one-click chips, or type any model name your provider supports. Leave it empty to use Agentum’s built-in default for that provider. The badge at the top of the card shows what’s live — e.g. Connected · Claude · claude-sonnet-4-6.

Connecting your own key

Use this if you already have API billing with a model provider, or want that provider to bill you directly. A consumer chat subscription does not necessarily include API usage — check the provider’s API terms first. If you don’t need that, the built-in option requires no separate account.

  1. Set up API billing and get a key from your provider. Claude (Anthropic), Gemini (Google), MiniMax or DeepSeek. Each has an “API keys” page in its developer account. A key is a long string of characters — treat it like a password.
  2. Open Settings → Connections. The LLM Provider card is the first one on that tab.
  3. Change “Active provider” to your provider. Only the one you pick needs credentials — the others can stay empty.
  4. Paste the key into the field that appears, and Save. There’s a reveal toggle if you want to double-check it. The badge in the corner updates to show your provider and model. That company now bills you directly for what your agents use.
  5. Click Restart agent when the banner asks — the running agents pick up the new provider on restart.
  6. Optionally, set a daily usage cap. “Daily usage cap per agent” limits what any single agent can spend per day, across every provider including Agentum AI. An agent that hits the cap pauses until midnight Pacific.

Switching back is the same three clicks. Nothing you’ve built is tied to a provider — your agents don’t know or care which AI is behind them, so changing it is a settings change, not a migration. Your other providers’ keys stay saved, so you can flip back and forth freely.

One caveat: the agent-building flows (Create Agent’s structured planning calls) need a provider that supports structured tool calls — that’s Agentum AI, Claude, Gemini, or DeepSeek. If you run MiniMax as your chat provider, Agentum tells you when a build needs one of the others.

The coding assistant

When Agentum builds or edits agent code, it doesn’t do it with the chat provider — it hands the work to a real agentic coding CLI running on your Mac, with the agent’s source folder and a prepared task.

Agentum Code (built-in) is the default, and like Agentum AI it runs on your plan’s monthly usage: no account to open, no key to paste. The managed model it uses is the one selected on the same card.

You can point it at a CLI you already pay for instead:

  Claude Code Codex
Made by Anthropic OpenAI
CLI binary claude codex
Sign-in Claude subscription or Anthropic API key ChatGPT account
Install curl -fsSL https://claude.ai/install.sh \| bash curl -fsSL https://chatgpt.com/codex/install.sh \| sh
Docs docs.claude.com developers.openai.com/codex/cli
Desktop-app hand-off Yes (Claude desktop app) No

Select yours under Coding assistant on the LLM Provider card. The setting is independent of the chat provider above, and applies to newly started sessions — a coding session that’s already open keeps whatever it launched with.

Detection and one-click install

You don’t have to pre-install anything. Agentum finds the CLI by probing your shell’s PATH plus the common install locations (~/.local/bin, Homebrew, npm, Volta, Bun…). If it’s missing when a session starts:

  • The inline session (embedded terminal) explains what’s missing, asks for consent, runs the official installer right there, and drops you into the CLI when it finishes.
  • The Terminal hand-off prints the exact install one-liner so you can run it yourself.

Where the coding assistant shows up

Three surfaces use the coding assistant — all covered in depth in Improving Agents:

  • Create Agent — the “draft it, then perfect it” hand-off after Agentum generates your first working version.
  • Improve — the per-agent tab that turns “add a snooze button” into a prepared coding session.
  • Sidekick — the companion rail, which can host a live inline coding session next to the agent.

Inline vs. Terminal vs. Desktop app

Every hand-off offers up to three places to run the session:

  • Inline — an embedded terminal inside Agentum itself. No window juggling; the session can move between the Improve tab, the Sidekick rail, and a full-screen view without restarting. This is the recommended default, and the only surface with the install-on-missing flow.
  • Terminal — opens macOS Terminal (or iTerm, auto-detected) with the CLI already launched in the agent’s folder and your task loaded. Good if you like your own terminal setup. Optional toggles: skip the CLI’s permission prompts, and auto-close the window when the session ends.
  • Desktop app — opens the task in the Claude desktop app via deep link. Claude Code only — the option is hidden when Codex is selected.

There’s also a fourth, tool-agnostic path: “Use another AI” copies a self-contained prompt to your clipboard for ChatGPT, Cursor, or anything else.

Frequently asked

Is Agentum AI included in my plan? Yes — we pay the model provider, and your plan’s monthly usage covers it. You never get a separate AI bill from us or from them. If you want headroom you control, connect your own key instead.

What if I use up my month? Alerts land at 80% and 95%; at 100% your agents pause rather than spending. From there you can reset the month (a full month’s usage lands on top of what’s left), move up a plan, or connect your own key in Settings → Connections — Agentum picks up where it left off either way, with your settings and agents unaffected. See Pricing for what a reset costs.

Does the coding assistant cost extra? Agentum Code runs on the same plan usage as everything else. If you switch to Claude Code or Codex, that CLI uses your own Anthropic/OpenAI account — a Claude subscription, an API key, or a ChatGPT plan, billed by them.

Can I use different tools for different agents? The coding-assistant setting is global, but it only affects how a session launches — you can switch it between sessions any time.

Is my chat provider used for coding at all? For the big code-writing sessions, no — that’s the coding assistant. Agentum’s own orchestration (planning your agent, reviewing what changed) uses the chat-side providers.

Next steps