Glossary
The vocabulary of an Anouk extension.
Short definitions for the terms you'll meet in the docs and templates.
- AIService
- The runtime entry point. AIService.call(instruction, content, requestId, cacheKey?) resolves config, checks the cache, queues, dispatches to the provider, decodes, caches, and returns a plain string.
- anouk init
- The CLI command that scaffolds a new Manifest V3 extension with the AI service, config manager, and settings panel pre-wired and runnable.
- anouk generate
- CLI command to drop an extension, service, or config template into an existing project: anouk generate extension|service|config.
- cache key
- The combination ${requestId}_${cacheKey} under which a response is stored. Identical instruction/content for the same key returns the cached answer.
- chrome.storage.local
- The browser storage area where the settings panel persists providerUrl, apiKey, model, and systemPrompt.
- configManager
- The module that resolves configuration, either from the runtime settings panel or a code-supplied aiConfig.js default export.
- content script
- The extension script that runs in the context of a web page. A common place to call AIService — e.g. to summarize or rewrite page content.
- Manifest V3
- The current Chrome extension platform. Anouk's templates target MV3 with host_permissions, storage, and unlimitedStorage.
- OpenAI-compatible
- Any endpoint that accepts the OpenAI chat-completions request shape. Anouk speaks this one shape so a single adapter reaches many providers.
- providerUrl
- The endpoint AIService POSTs to. Set it to a hosted provider, a local Ollama server, or even your own backend.
- rate-limit queue
- A per-provider queue that serializes outbound calls so bursts of user activity do not stampede the provider or trigger 429 responses.
- settings panel
- The bundled runtime UI where the end user enters their provider URL, key, model, and system prompt without a rebuild.
- systemPrompt
- A configurable instruction sent with each request to steer the model. Part of the settings panel and aiConfig fields.