> ## Documentation Index > Fetch the complete documentation index at: https://docs.agentmark.co/llms.txt > Use this file to discover all available pages before exploring further. # Generation types > Understand the different types of content you can generate with AgentMark prompts Generation types define what kind of output your prompt produces. AgentMark supports four types, each suited to different use cases: * **Text**: natural language responses for chatbots, content generation, and analysis * **Object**: structured JSON data with schema validation for APIs and data extraction * **Image**: visual content from models like DALL-E 3 * **Speech**: spoken audio for voice applications and text-to-speech ## Choosing the right type | Type | Best for | Output format | Example use cases | | ---------- | ---------------------------------- | ---------------- | ----------------------------------------------- | | **Text** | Conversational AI, content writing | String | Chatbots, summarization, Q\&A | | **Object** | Structured data extraction | JSON with schema | Form parsing, data normalization, API responses | | **Image** | Visual content creation | Image file | Marketing assets, illustrations, prototypes | | **Speech** | Voice applications | Audio file | Podcasts, audiobooks, voice assistants | You declare each type with its frontmatter config key (`text_config`, `object_config`, `image_config`, or `speech_config`) and load it with the matching client method (`loadTextPrompt`, `loadObjectPrompt`, `loadImagePrompt`, `loadSpeechPrompt`). This page covers text, image, and speech; [object generation](/build/generation-types/object) has its own page for schema validation and `$ref` reuse. ## Text generation AgentMark generates text with prompts that declare a `text_config` in frontmatter. Text prompts use message-role tags (``, ``, ``) and return a string. ### Example configuration ```jsx example.prompt.mdx theme={null} --- name: example text_config: model_name: openai/gpt-5-mini --- You are a math tutor that can perform calculations. What's 235 * 18? ``` These docs write `model_name` in the provider-prefixed form, such as `openai/gpt-5-mini`. The value is a free-form string that AgentMark passes through to your call site unchanged, which is why the [Running prompts](/build/running-prompts) examples strip the `openai/` prefix before constructing the provider model. ### Tags | Tag | Description | | ------------- | --------------------------------------------------------------------------------- | | `` | System-level instructions | | `` | User message | | `` | Assistant message (optional; include for few-shot examples or prior-turn context) | ### Available configuration | Property | Type | Description | Required | | ------------------- | ----------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------- | | `model_name` | `string` | The name of the model to use for text generation. | Yes | | `max_tokens` | `number` | Maximum number of tokens to generate. | No | | `temperature` | `number` | Controls the randomness of the output; higher values are more random. | No | | `max_calls` | `number` | Maximum number of LLM calls allowed (for agent workflows). | No | | `top_p` | `number` | Cumulative probability for nucleus sampling. | No | | `top_k` | `number` | Limits next-token selection to the top `k` tokens. | No | | `presence_penalty` | `number` | Penalizes tokens based on presence in the text so far, encouraging new topics. | No | | `frequency_penalty` | `number` | Penalizes tokens based on frequency in the text so far, reducing verbatim repetition. | No | | `stop_sequences` | `string[]` | Strings that, if encountered, stop generation. | No | | `seed` | `number` | Random-number seed for reproducibility. | No | | `max_retries` | `number` | Maximum number of retries on failure. | No | | `tool_choice` | `"auto" \| "none" \| "required" \| { type: "tool", tool_name: string }` | Controls how the model uses tools during generation. | No | | `tools` | `string[]` | List of tool names or MCP URIs available to the model. You resolve these names to implementations at your call site (see [Tools and agents](/build/tools-and-agents)). | No | ### Running a text prompt See the **SDK usage** section of [Running prompts](/build/running-prompts) (under the **Local** tab) for the text-generation SDK code patterns: render the prompt with `prompt.format()`, then make your own model call (`generateText` in TypeScript, the OpenAI client in Python). ## Image generation AgentMark generates images with prompts that declare `image_config` in frontmatter. The image description itself goes in an `` tag. ### Example configuration ```jsx example.prompt.mdx theme={null} --- name: image image_config: model_name: openai/dall-e-3 num_images: 1 size: 1024x1024 aspect_ratio: 1:1 seed: 12345 --- A futuristic cityscape at sunset with flying cars and neon lights ``` ### Tags | Tag | Description | | --------------- | -------------------------------------------------------------------------------------------------------------------------------- | | `` | The text description for image generation. AgentMark reads the contents at compile time and sends it to the model as the prompt. | ### Available configuration | Property | Type | Description | Required | | -------------- | -------- | -------------------------------------------------------------------------------- | -------- | | `model_name` | `string` | The name of the model to use for image generation. | Yes | | `num_images` | `number` | Number of images to generate. | No | | `size` | `string` | Image dimensions in format `WIDTHxHEIGHT` (for example, `1024x1024`, `512x512`). | No | | `aspect_ratio` | `string` | Aspect ratio in format `WIDTH:HEIGHT` (for example, `1:1`, `16:9`, `9:16`). | No | | `seed` | `number` | Random-number seed for reproducibility. | No | Set `size` for pixel-exact dimensions or `aspect_ratio` for proportional sizing. AgentMark passes both through to the model unchanged, so provider support varies. If you set both, the provider chooses which to honor. ### Running an image prompt See the **SDK usage** section of [Running prompts](/build/running-prompts) (under the **Local** tab) for the image-generation SDK code pattern using Vercel AI SDK's `experimental_generateImage`. ### Tracing image generation AgentMark's prompt runner (deployed agents and experiments) captures the generated image on the span automatically, and the trace's Output tab [renders it inline](/observe/traces-and-logs#generated-media). Your own application code has a gap. The Vercel AI SDK's `experimental_generateImage` (unlike `generateText`) emits no OpenTelemetry telemetry, so tracing captures nothing on its own. Instrument the call yourself and set the generated media as the span output: ```ts theme={null} import { span } from "@agentmark-ai/sdk"; import { experimental_generateImage as generateImage } from "ai"; import { openai } from "@ai-sdk/openai"; const { result } = span({ name: "generate-image", kind: "llm" }, async (ctx) => { ctx.setInput({ prompt: "a red sneaker on a white background" }); const { images } = await generateImage({ model: openai.image("gpt-image-1"), prompt: "a red sneaker on a white background", }); // Set the media as the output so it renders inline in the trace. ctx.setOutput(images.map((i) => ({ mediaType: i.mediaType, base64: i.base64 }))); return images; }); await result; ``` For the image to render, the output must be (or contain) objects shaped `{ mimeType, base64 }`. The `mediaType` key also works. A value of any other shape renders as text. The same pattern applies to audio from `experimental_generateSpeech`. See [Tracing setup](/observe/tracing-setup) for initializing the tracer. ## Speech generation AgentMark generates speech audio with prompts that declare `speech_config` in frontmatter. The text to speak goes in a `` tag. ### Example configuration ```jsx example.prompt.mdx theme={null} --- name: speech speech_config: model_name: openai/tts-1-hd voice: "nova" speed: 1.0 output_format: "mp3" --- Please read this text aloud. This is a test for the speech prompt to be spoken aloud. ``` ### Tags | Tag | Description | | ---------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `` | The text to convert to speech. AgentMark reads the contents at compile time and sends it to the TTS model. | | `` | Optional system-level instructions for models that support them. At compile time the tag body becomes `instructions` in the compiled `speech_config` (or `""` when the tag is absent). AgentMark ignores an `instructions` value written directly in frontmatter. | ### Available configuration | Property | Type | Description | Required | | --------------- | -------- | ---------------------------------------------------------------------------------------------------------------------- | -------- | | `model_name` | `string` | The name of the model to use for speech generation. | Yes | | `voice` | `string` | Voice identifier (provider-specific; for example for OpenAI TTS: `alloy`, `echo`, `fable`, `onyx`, `nova`, `shimmer`). | No | | `output_format` | `string` | Audio output format (for example, `mp3`, `opus`, `aac`, `flac`). | No | | `speed` | `number` | Playback speed multiplier. | No | The compiled `speech_config` also carries `text` (from the `` tag) and `instructions` (from the `` tag). AgentMark populates both at compile time from the tags; you don't author them in frontmatter. ### Running a speech prompt See the **SDK usage** section of [Running prompts](/build/running-prompts) (under the **Local** tab) for the speech-generation SDK code pattern using Vercel AI SDK's `experimental_generateSpeech`.

Have questions?

Reach out any time: