Why prompts don’t belong in your code
Most LLM features start with the prompt written inline, right next to the call to the model:- Every wording change is a code change: a pull request, a build, a deploy, all for a tweak a product manager could have made in a text box.
- There’s no history. You can’t see what the prompt said last Tuesday, who changed it, or why the quality dropped after the change.
- You can’t compare two versions without digging through commits.
- Only engineers can touch it, even though the people who understand the prompt best often aren’t engineers.
How to11 models a prompt
A prompt is a named, versioned template that a project owns, alongside that project’s skills, environments, and sessions. It has a stable slug (weather-concierge) your application renders by, and a history of versions behind it.
There are two kinds of prompt. This page describes a chat prompt, which your application renders into messages and sends to a model. The other kind is a skill — a folder of files delivered to a developer’s machine, where a coding agent reads it. Both use the same versions, the same
latest label and the same history; what differs is what they carry and who they are handed to. A prompt’s kind is chosen when it is created and never changes. Skills have their own section, starting at Skills.
Authoring happens on a draft. Publishing freezes the draft into a read-only version, moves the prompt’s one label,
latest, to it, and starts a fresh draft from the same blocks, so what’s live never changes underneath you. See Versions.
Message roles
Each block is a message with a role that tells the model who is speaking and how much authority the message carries.system and developer are the author’s roles, and they outrank the end user’s user messages, so a rule you set isn’t overridden by what a user types. assistant blocks and tool-result blocks stand in for earlier turns in the conversation. The Authoring page describes what each role is for; how a developer block reaches a model that doesn’t accept that role is covered under Rendering.
to11 serves the prompt; your code sends the call
to11 keeps the prompt and hands it to your application; it is not part of the model call itself:- to11 (the dashboard and the SDK’s
prompts.*methods, backed by the to11 API) is where you author and publish versions, and where your application renders one. - Your provider client (the OpenAI or Anthropic SDK, or any other) sends the rendered messages, tools, and model settings to the provider, with your own provider key.
The lifecycle
- Author. Build the version’s blocks, variables, tools, and config in the editor. See Authoring.
- Publish. Freeze the draft into an immutable version;
latestmoves to it. See Versions. - Render. Your application renders
latest, which is always the newest version. See Rendering prompts.
Next steps
Authoring prompts
Blocks, variables, conditions, tool definitions, and model config.
Versions
Publish immutable versions;
latest follows the newest.Rendering prompts
Render a prompt’s newest version from your application at runtime.