Grok Integration

Grok runs xAI's models inside your workflows. You sign in once with the Grok account you already have, then use that session in AI nodes. There is no API key to create, paste, or rotate, and no paid plan to buy first.

Usage counts against your Grok account, not against Solaris AI. We never see your password.

What you get

Grok is one of four AI providers in Solaris AI. The others are ChatGPT (device login, your ChatGPT account), OpenRouter (hundreds of models, one key) and Venice AI (private inference). You can connect more than one and switch per node.

Connect it

Unlike ChatGPT, Grok needs no one-time setting on the provider's side. You can sign in straight away.

From Connections (recommended)

1. Open Connections in the dashboard sidebar.
2. Expand Grok under AI Models.
3. Click Add key, then read the short privacy note and click Continue with Grok.
4. Copy the one-time code (or leave it on screen).
5. Click Open xAI to enter the code and finish authorization on xAI's site with your Grok account.
6. Come back to the Solaris AI tab. It picks up the authorization on its own, and the dialog turns into Grok connected.

You should see a Connected badge on Grok. The credential is titled with the account that signed in, for example Grok (a•••@example.com). Multiple sessions are labeled as sessions, not keys.

From an AI node

Open an AI node on the canvas. If no AI credential is available, the config prompt offers Grok sign-in alongside ChatGPT, OpenRouter, and Venice. Connect there, pick the new credential, then choose a model.

Using an xAI API key instead

xAI runs an allowlist on its API surface, so a valid Grok login can still be refused API access with a 403. When that happens, an API key is the fix, and it is offered on the same dialog:

1. Create a key at console.x.ai. Usage is billed to your xAI account balance.
2. In the Grok connect dialog, click Have an xAI API key? Use it instead.
3. Paste the xai-... key and click Save API key.

Solaris AI validates the key against xAI before saving it, so a mistyped key is rejected up front rather than at run time. The credential then behaves like any other Grok credential: same node picker, same models, same outputs.

Reconnecting an expired session

Grok sessions can expire. When that happens:

• Connections shows Session expired on that row.
• A run may say the Grok session expired.
• The dashboard raises "Your Grok session expired, workflows using it will fail".

Click the pencil on the expired row and complete the device flow again. Reconnecting into the same credential keeps every AI node that already selected it working, so you do not need to rewire workflows.

Use it

Drop an AI node, set AI Provider Key to your Grok credential, pick a model from the live dropdown, and write a prompt. Downstream nodes read the reply as {aiResponse.data}.

The point of the AI node is not to write text. It is to put judgment between the data you pull and the action you take. Solaris AI Flow exposes more than 200 onchain operations across Solana protocols and data providers, including over 100 DeFi actions on Jupiter, Raydium, Orca, Meteora, Kamino, Phoenix, Sanctum, DFlow, and Pump.fun. The model reads the data those nodes return and decides what happens next.

Typical uses:

• Gate a swap on conditions. Pull a Jupiter quote and a Pyth price, ask the model whether the spread and slippage justify the trade, and let a Condition node route to a Jupiter swap or to nothing.
• Watch leverage. Read getUserObligations and getLeverageMetrics from Kamino, have the model judge how close a position sits to liquidation, and alert on Telegram before it matters.
• Rank what's new. Feed trending tokens from GMGN or Pump.fun into the model, score them against your own criteria, and place a Phoenix limit order on the survivors.
• Reason over a long haul. Grok 4.5 takes a 500K-token context, so a whole run's worth of raw data can go in at once instead of being summarized first.

Set responseFormat: json_object when a downstream node needs to branch on specific fields rather than read prose.

Full node reference: AI. If a workflow template or credential gate asks for an AI provider, Grok satisfies the same requirement as ChatGPT, OpenRouter, or Venice for text models.

Models

xAI's lineup is currently a single model, and the dropdown is live from your account, so what you see is what you can run.

Model IDs are bare, like grok-4.5, not slashed the way OpenRouter's are. You can type an ID by hand, and a model xAI ships after this page was written still works: the node sends it as-is and reads the live list for what your account can reach.

Reasoning

Grok 4.5 always reasons. It accepts low, medium, or high effort, and defaults to high. There is no way to switch reasoning off.

The node maps whatever you pick onto what the model actually accepts instead of sending a value it would reject:

• Turning Reasoning off does not stop it reasoning. The effort is simply left unset and the model uses its own default (high).
• Extra High and Max are lowered to high, the deepest Grok 4.5 takes. Minimal is raised to low.

Reasoning output lands in the reasoning field alongside the reply, the same as OpenRouter and Venice.

How it works

1. You start Continue with Grok in Solaris AI.
2. Solaris AI asks xAI for a short one-time code and a verification link.
3. You open xAI's page, enter the code, and approve access with your Grok account.
4. Solaris AI stores an encrypted session (not your password).
5. When you run an AI node, the runtime uses that session to call models on your account.

Click Connect → one-time code → authorize on xAI → encrypted session → AI node runs

While you are on xAI's page, Solaris AI keeps polling for the result, so the dialog flips to connected on its own once you approve. It survives a backgrounded tab or a brief network drop.

Nothing sensitive is shown in the browser except the code you type on xAI's site. The session is decrypted only at run time, the same way other credentials work. See Security for encryption details.

Grok vs ChatGPT vs OpenRouter vs Venice

Use Grok when you already have a Grok account and want it to do the work. Switch to OpenRouter when you need a specific commercial model, free-tier options, or image generation.

Limits and good to know

• Text only in Solaris AI. Image output requires OpenRouter.
• Session can expire. Reconnect from Connections; your workflows keep the same credential id.
• xAI allowlists its API. A valid login can still be refused API access with a 403. The fix is an xAI API key on the same credential.
• Models depend on your account. The dropdown is live from your session. xAI retires models over time, and a retired slug does not fail loudly: it redirects to whatever replaced it and bills at that model's rate. So a model saved on an old node can quietly become a different model. Pick a current one from the dropdown.
• Reasoning cannot be turned off, only turned down. See Reasoning.
• Penalties and stop sequences are dropped, because every Grok model reasons. See Reasoning.
• No password sharing. Authorization always happens on xAI's domain with a one-time code.
• Marketplace clones never include your Grok session. Buyers connect their own account.

Troubleshooting

Privacy and security (short version)

• You authorize on xAI's site; Solaris AI never collects your X password.
• The session is encrypted at rest (AES-256-GCM with Cloud KMS), same as other credentials.
• Tokens are used only for runs you start.
• Credentials are not exported with workflows or marketplace listings.

More detail: Security and Credentials.

Next steps

• Connect other providers: ChatGPT, OpenRouter, Venice, data APIs, messaging
• AI node: prompts, JSON mode, templates, and outputs
• Node reference: every swap, order, and lending action an AI node can gate
• Copilot: build workflows from a prompt (not live yet)
• Per-integration guide: where to get keys for every platform