xAI Grok
AIRun xAI's frontier Grok models: chat completion, real-time answers via Live Search across web, X, and news, vision, image and video generation, and structured output.
What This Integration Enables
Grok pairs frontier reasoning with real-time grounding. Chat completion and structured output handle generation, vision reads images, image and video generation produce assets, and Live Search grounds answers in current information across web, X, and news. For anything where recency matters, the live grounding is what makes the answer useful. Real-time awareness is only an advantage if the flow acts on it fast and correctly. Knowing a brand mention is spiking is a prompt for a decision, not a decision. An orchestration layer gathers the context, summarizes it with sources, and routes it to the team who owns the response, pulling a person in before anything public happens. FlowRunner is built for that layer.
Without FlowRunner
With FlowRunner
Use Case Scenarios
Real-Time Brand Monitoring
Monitoring picks up a spike in brand mentions. The agent runs Live Search across web, X, and news, summarizes what is driving the spike with sources, and gauges the sentiment and reach. The comms team gets a sourced situation summary within minutes instead of piecing it together from scrolling. Any summary suggesting a public response routes to comms to decide.
Current-Events Q&A
A workflow needs answers grounded in what is happening now. The agent uses Grok with Live Search to answer from current sources with references. Internal answers return directly; anything that will inform a public statement is routed to a person to verify before it is used.
Multimodal Content Generation
A campaign needs on-brand images and short video from a brief. The agent generates candidates with Grok's image and video generation. The drafts go to a person for selection and approval before publishing, so generation accelerates the work without removing creative and brand judgment.
Human-in-Loop Highlight
Real-time context is exactly the kind of input that tempts a fast, public reaction, which is why a human gate belongs in front of it. When Grok's Live Search surfaces a situation that might warrant a public response, or generates content headed for publication, FlowRunner routes it through a [human-in-the-loop](/concepts/human-in-the-loop/) step: the agent pauses, presents the summary with its sources, and sends it to the comms team via Slack. They decide whether and how to respond. The model delivers the real-time picture; a person owns the public move.
Agent Capabilities
18 actionsChat
3- Chat Completion Generates a text response for a single prompt using an xAI Grok model (Grok 4. 5, Grok 4.
- Chat Completion (Advanced) Sends a fully custom chat completion request to xAI with a complete messages array (multi-turn conversations, multimodal image_url content parts), tool/function calling passthrough, structured outputs via a response format object (json_object or json_schema), Live Search parameters, reasoning effort, penalties and sampling controls.
- Get Deferred Completion Retrieves the result of a deferred chat completion by its request ID (obtained from 'Chat Completion (Advanced)' with Deferred enabled).
Live Search
1- Ask with Live Search Answers a prompt with real-time information using xAI's Live Search: the Grok model searches live data sources (web, X posts, news, RSS feeds) and grounds its answer in what it finds.
Vision
1- Analyze Image Analyzes one or more images with a vision-capable Grok model and answers a prompt about them, describe content, extract text, compare images or answer visual questions.
Image Generation
2- Generate Image Generates one or more images from a text prompt using xAI's Grok Imagine image models. Supports aspect ratio and resolution control.
- Edit Image Edits an existing image (or combines multiple reference images) with natural-language instructions using xAI's Grok Imagine models, change styles, add or remove objects, or merge content from several references.
Video Generation
4- Generate Video Starts an asynchronous video generation with xAI's Grok Imagine video models from a text prompt and an optional starting image (image-to-video).
- Edit Video Starts an asynchronous edit of an existing video using natural-language instructions with xAI's Grok Imagine video models, restyle scenes, change elements or adjust the mood.
- Extend Video Starts an asynchronous extension of an existing video, generating a continuation described by the prompt with xAI's Grok Imagine video models.
- Get Video Result Retrieves the result of an asynchronous video generation, edit or extension request by its request ID.
Utilities
1- Tokenize Text Tokenizes text with a specified Grok model's tokenizer and returns the token list and token count.
Account
1- Get API Key Info Retrieves information about the configured xAI API key, including its name, redacted value, access control list (permitted endpoints and models), and whether the key or team is blocked or disabled.
Models
5- List Models Lists all xAI models available to the configured API key across every modality (chat, image understanding, image generation and video generation) with basic information.
- Get Model Retrieves basic information about a specific xAI model by its ID, including its owner, creation date and pricing basics.
- List Language Models Lists all xAI Grok language models (chat and image understanding) with full metadata: context window, aliases, input/output modalities, per-token prompt/completion/cached pricing and Live Search pricing.
- List Image Generation Models Lists all xAI image generation models (Grok Imagine) with full metadata: aliases, modalities, per-image pricing and maximum prompt length.
- List Video Generation Models Lists all xAI video generation models (Grok Imagine video) with full metadata: aliases, modalities and pricing.
Start building with xAI Grok
$100 in credits. No card required. Connect in minutes.