LLM
Managed text generation and project Knowledge for your app.
Generate, rewrite, summarize, translate, classify, or extract text with GPT-5.6 Luna. Add Knowledge when answers should come from your project's own documents.
Your coding agent enables the service and adds the right app connection. No separate model account or provider key is required.
Text and Knowledge
- Text: send input with optional instructions for tone, language, or format; receive generated text and token usage.
- Multilingual: use Korean, English, Japanese, Spanish, French, Chinese, German, and other languages.
- Knowledge: link files already in Orchard Storage, then receive answers with the source filenames used.
- Privacy: Orchard records aggregate usage, not prompts or generated answers.
Add project Knowledge
Each project has an isolated Knowledge index. Orchard prepares the linked documents for search, finds relevant passages, and keeps one project's sources separate from every other project.
- Upload documents with Orchard Storage.
- Ask your agent to enable Knowledge and link the filenames.
- When indexing is ready, send LLM requests with Knowledge enabled.
Linked files can be up to 4 MB and include common document, text, spreadsheet, data, and code formats. Removing the Storage source also removes its linked Knowledge copy.
Connect your app
Apps with a backend call Orchard with a server-side project key.
Never place that key in browser code. An Orchard Deploy site can
instead use the same-origin /_orchard/llm path after
the Builder enables deployed-site access.
Public deployed-site calls count toward the account's included usage and spending cap. The Builder owns app login, rate limits, and any additional abuse controls.
Current boundaries and pricing
- LLM accepts text only. It does not search the live web, generate images, call tools, enforce a JSON schema, or remember conversation history automatically.
- Each request supports up to 20,000 input characters, 4,000 instruction characters, and 4,096 output tokens.
- Free includes $0.20 of LLM processing per billing period; Builder includes $0.40, then usage is metered.
- View token usage, pause generation, and manage the account spending cap from the dashboard.
See current included usage and overage rates on Pricing.
Ask your agent
- "Add a server-side LLM reply feature to this support form."
- "Make a support assistant that answers from these product PDFs and shows its sources."
- "Let this Orchard Deploy site use LLM, then show me the usage controls."