Generate Text

Overview#

Generate text with GPT, Claude, or Gemini.

Use this block when its responsibility is clear in the graph. Give it a name that describes the work it performs, then test it with representative input before publishing.

Working examples#

These are complete starting points, not decorative diagrams. Every graph is also available in the main template library and is checked by the workflow compiler.

1. Meeting notes to decisions#

You will build: Owners, decisions, deadlines, and unresolved questions.

A focused extraction prompt is more dependable than a generic summary. A bounded output format makes missing owners and dates visible.

Open this template and adapt the labels, prompts, or credentials to your own workflow.

2. Policy plain-language rewrite#

You will build: A faithful rewrite that preserves obligations.

Separates clarity work from legal review. The prompt instructs the model not to invent or remove requirements.

Open this template and adapt the labels, prompts, or credentials to your own workflow.

3. Lead qualification memo#

You will build: Fit signals, risks, and the next discovery question.

Turns inconsistent notes into a repeatable sales handoff. Explicit evidence fields keep the assessment traceable to the input.

Open this template and adapt the labels, prompts, or credentials to your own workflow.

Ports#

• Input: Input (in): any, optional, single.
• Output: Response (out): string, optional, single.

Inspector#

• Prompt: key prompt; textarea; default: "".
• Style / System (optional): key system; textarea; default: "".
• Model: key model; select; options: gpt-6-astra, gpt-6-sol, gpt-6-luna, Kimi-K2.6, DeepSeek-V4-Pro, claude-sonnet-5, claude-opus-5-5, gemini-3.7-flash, gemini-3.1-pro-preview, gemini-3.5-flash-lite, gemini-3.1-flash-lite; default: "gpt-6-sol".
• Max Tokens: key maxTokens; number; range: 1 to 16000; default: 2000.
• Timeout (ms): key timeout; number; range: 30000 to 600000; default: 120000.
• Stream: key stream; inline switch; default: true.
• expanded-text editor: manages prompt, system with edit, expand support.

Behavior#

• Input requirement: Provide one of ports in or config fields prompt, system.
• Execution: Handled by the handler using llm-chat.
• Timeout: 120000 ms by default (30000 to 600000 ms).

Cost#

Uses a language-model provider and contributes model usage to the run cost.

Security#

• Uses a runner-provided model API key. Secret values are not part of the published graph.
• The node reads from an external service without intentionally mutating it.

Retry#

The runtime may retry this node according to the compiled execution policy.

Production checklist#

• Test the happy path and one malformed or empty input.
• Keep credentials out of prompts and published graph configuration.
• Set a timeout and retry policy that matches the operation's side effects.
• Name the final result so a runner can tell what success looks like.