Get an answer from Claude Haiku 4.5

Give your Agent a task, a brief role instruction, 1–2 prior messages and a web-search country. It requests a Claude Haiku 4.5 response through DataForSEO and presents available answer text and references.

Your AI

Use with your AI

For tasks with several steps, continue in the AI you already use.

CodexClaude CodeCursorWindsurfCline

View instructions

Use the exact Toolmu Tool claude-llm-responses-live to answer the user’s task. Reuse the existing Toolmu connection; if needed, follow https://toolmu.com/SKILL.md. Work within the user’s current authorization and budget. Ask only for missing required context: user_prompt, system_message, 1–2 actual prior message_chain turns with role and message, and a supported web_search_country_iso_code such as FR, US or GB. Keep every submitted text value within 1–256 characters; use the supplier’s user/ai roles for supplied history and never invent prior turns. Send all eight request properties, setting model_name to claude-haiku-4-5, web_search to true, temperature to 0.3 and max_output_tokens to 200. Present available message text and any source titles/URLs; retain missing values and report search use only when returned. Treat the 200-token setting as a request that can be exceeded, and distinguish the UTC receipt time from source freshness. Do not imply that this continues the user’s Claude account session or independently verifies the answer.

Task:
Answer [your question] using [why you are asking] as background. Reply in [language] with [a concise, practical explanation], and look for information online about [country] when possible.

Paste into your connected AI's conversation.

Connection guide

Up to 0.11 credits / call

The final charge is based on confirmed usage. Unused reserved credits return to your available balance.

DataForSEO

Get an answer from Claude Haiku 4.5

You receive

Receive one Claude Haiku 4.5 response for the task and context you provide. Your Agent can organize the available text sections into an answer for you to review, with any returned source titles and URLs alongside it. Reported model, token counts, search use and the UTC receipt time may help you understand the response.

Start with your task

Answer [your question] using [why you are asking] as background. Reply in [language] with [a concise, practical explanation], and look for information online about [country] when possible.

What to know

The model is fixed and web search is enabled, but actual searching, references and current facts are not guaranteed. The requested 200-token setting can be exceeded; it is not a 200-word limit. Supply your own prior messages rather than expecting a Claude account conversation to carry over.

Inputs & outputs

Inspect the current inputs, availability, and maximum charge before starting.

Input schema

Input schema
Field / Path	Type	Requirement	Meaning
Response request`request`	`object`	Required	All eight settings are required: the current task, role instruction, supplied history, search country and four fixed model settings. Required fields `system_message, message_chain, max_output_tokens, model_name, temperature, web_search, web_search_country_iso_code, user_prompt`
Claude model`request.model_name`	`string`	Required	Fixed to claude-haiku-4-5, the Haiku 4.5 model identifier. This Tool does not offer model selection. Fixed value `claude-haiku-4-5`
Web search enabled`request.web_search`	`boolean`	Required	Must be true. Web retrieval is enabled, but actual search use and source references depend on the response. Fixed value `true`
Response temperature`request.temperature`	`number`	Required	Fixed to 0.3. This setting influences sampling; it does not make repeated answers identical. Fixed value `0.3`
Current task`request.user_prompt`	`string`	Required	The question or task to answer now. Supply 1–256 characters. Minimum length `1` Maximum length `256`
Prior messages`request.message_chain`	`array`	Required	Supply 1–2 actual previous messages. Each message has a role and message text; account conversation history is not imported automatically. Minimum items `1` Maximum items `2`
Prior message`request.message_chain[]`	`object`	—	One earlier turn supplied as an object containing both role and message. Required fields `role, message`
Message role`request.message_chain[].role`	`string`	—	Role of the earlier speaker. The supplier documents user for the user and ai for the model; submit 1–256 characters. Minimum length `1` Maximum length `256`
Message text`request.message_chain[].message`	`string`	—	Text of the earlier turn, 1–256 characters. Include only history the user has supplied. Minimum length `1` Maximum length `256`
Role and tone instruction`request.system_message`	`string`	Required	A brief instruction for the model’s role, tone or behavior, 1–256 characters. Minimum length `1` Maximum length `256`
Requested token limit`request.max_output_tokens`	`integer`	Required	Must be 200. Tokens are not words; the supplier allows the generated count to exceed this setting when web search is enabled. Fixed value `200`
Web-search country`request.web_search_country_iso_code`	`string`	Required	A supported ISO country code for search context, such as FR, US or GB. Submit 1–256 characters; these examples are not defaults or city selectors. Minimum length `1` Maximum length `256`

View raw JSON

{
  "type": "object",
  "required": [
    "request"
  ],
  "properties": {
    "request": {
      "type": "object",
      "required": [
        "system_message",
        "message_chain",
        "max_output_tokens",
        "model_name",
        "temperature",
        "web_search",
        "web_search_country_iso_code",
        "user_prompt"
      ],
      "properties": {
        "model_name": {
          "type": "string",
          "const": "claude-haiku-4-5"
        },
        "web_search": {
          "type": "boolean",
          "const": true
        },
        "temperature": {
          "type": "number",
          "const": 0.3
        },
        "user_prompt": {
          "type": "string",
          "maxLength": 256,
          "minLength": 1
        },
        "message_chain": {
          "type": "array",
          "items": {
            "type": "object",
            "required": [
              "role",
              "message"
            ],
            "properties": {
              "role": {
                "type": "string",
                "maxLength": 256,
                "minLength": 1
              },
              "message": {
                "type": "string",
                "maxLength": 256,
                "minLength": 1
              }
            },
            "additionalProperties": false
          },
          "maxItems": 2,
          "minItems": 1
        },
        "system_message": {
          "type": "string",
          "maxLength": 256,
          "minLength": 1
        },
        "max_output_tokens": {
          "type": "integer",
          "const": 200
        },
        "web_search_country_iso_code": {
          "type": "string",
          "maxLength": 256,
          "minLength": 1
        }
      },
      "additionalProperties": false
    }
  },
  "additionalProperties": false
}

Result schema

Output fields have not been documented yet.

Additional notes

For the supplier’s response and reference definitions, see DataForSEO Claude LLM Responses documentation.

Questions about this tool.

Inputs, results and details to know before running it.

Get connected

This Tool uses Haiku 4.5 with web search enabled, temperature 0.3 and a requested 200-token limit. These settings are fixed; it has no model or reasoning selector.

Get an answer from Claude Haiku 4.5

Get an answer from Claude Haiku 4.5

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Share your question and background

Choose a language and country

Check the answer and sources

Questions about this tool.

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Share your question and background

Choose a language and country

Check the answer and sources

Questions about this tool.

Can I choose a different Claude model or settings?

What history do I need to provide?

Will every answer include current information and citations?

Does 200 tokens mean a 200-word answer?

How should I read the returned response?