AI provider

Several Norish features are powered by AI. They're off by default, configure a provider to enable them.

AI enables:

• AI fallback when a recipe can't be imported from a URL structurally
• Image import from screenshots or photos of recipes
• Video import from YouTube Shorts, Instagram Reels, TikTok, Pinterest, and more
• Recipe Enrichment: tags, allergy indications, meal categories, nutrition values and ingredient to step linking.
• Unit conversion between metric and US units

Enable AI via the environment

note

AI can also be enabled via the admin settings.

Set AI_ENABLED=true and configure a provider. Norish speaks the OpenAI API format, so any OpenAI-compatible endpoint works (OpenAI, Azure OpenAI, Open Router, a local Ollama/LM Studio server, …).

docker-compose.yml (environment)

AI_ENABLED: "true"
AI_PROVIDER: openai
AI_MODEL: gpt-5-mini
AI_API_KEY: <your-api-key>
# For an OpenAI-compatible endpoint (Azure, OpenRouter, Ollama, …):
# AI_ENDPOINT: https://your-endpoint/v1

Variable	Description	Default
AI_ENABLED	Enable AI features globally	false
AI_PROVIDER	AI provider	openai
AI_ENDPOINT	Custom OpenAI-compatible endpoint	(empty)
AI_MODEL	Default model	gpt-5-mini
AI_API_KEY	API key for the provider	(empty)
AI_TEMPERATURE	Generation temperature	1.0
AI_MAX_TOKENS	Maximum tokens for model responses	10000
AI_TIMEOUT_MS	Maximum time for an AI response (ms)	300000

note

AI feature speed and quality vary by provider, model, and region. You can also adjust AI settings at runtime in Settings => Admin.

Recipe Enrichment

Recipe Enrichment is the optional AI work that runs after a recipe is saved: auto-tagging, allergy detection, auto-categorization, nutrition estimation, recipe provenance, and ingredient linking.

Importing and creating a recipe never depend on it. The recipe is saved first; enrichment is enrolled separately, and a disabled, unavailable, slow, or failing AI provider cannot make a save fail.

Automatic enrichment

Under Settings => Admin => AI, each kind has its own switch. They apply to every newly created recipe, manual entry and every import path alike.

Switch	What it does automatically	Default
Auto-tagging	Adds suggested tags without removing existing ones	Off
Allergy detection	Adds allergy tags for your household's configured allergies	On
Auto-categorization	Sets meal categories on recipes that have none	Off
Nutrition estimation	Estimates calories, fat, carbs, and protein when the recipe doesn't already have all four	Off
Recipe Provenance	Works out the country, region, cuisines, and a short note	Off
Ingredient Linking	Links ingredient lines to the steps that have none	Off

Enabling AI globally does not switch these on by itself, each is opt-in (except allergy detection, which keeps the behaviour of the setting it replaced). Turning one off only stops the automatic run; household members can still request that kind by hand from the recipe.

Automatic enrichment runs once, when a recipe is first created. Editing a recipe later does not trigger it again, and a URL import that matches a recipe you already have is not treated as a new recipe.

Supplied recipe data wins

Information you entered yourself, or that an import source stated explicitly, outranks automatic enrichment. Each group has its own precedence rule:

• Any meal category on the recipe suppresses automatic categorization.
• A complete nutrition group, calories, fat, carbs, and protein all present, suppresses automatic nutrition estimation; zeros count as present. An incomplete group does not: the estimate replaces the group as a whole, so the four values always agree with each other rather than mixing a supplied figure with an estimate.
• Any part of provenance, country, region, a cuisine, or the note, suppresses automatic provenance inference for the whole group. The note explains the whole claim, so it is never mixed with a value you set yourself.
• Ingredients are decided per step: a step you linked yourself is left alone, and only steps with no links at all are filled. This holds for a run you request by hand too. See Step ingredients.
• Empty and blank values do not count as supplied, so placeholders don't block useful enrichment.

Tags and allergy indications work differently: enrichment appends findings and never removes what is already there, so existing tags never suppress it.

A run you request by hand is a deliberate refresh and does replace the current categories, the complete nutrition group, or the complete provenance group.

Tag strategy

Tag strategy decides which tags auto-tagging may use, independently of whether it runs automatically:

Strategy	Behaviour
Predefined tags only	Only Norish's built-in tag list
Predefined + existing tags	Also tags already used by recipes on this instance
AI can create new tags	May invent new tags when nothing fits

Turning automatic auto-tagging off keeps the selected strategy for manual runs.

Cuisine strategy

Cuisine strategy decides whether provenance inference may add to the cuisine list your administrator maintains, independently of whether it runs automatically:

Strategy	Behaviour
Only existing cuisines	Pick from the list; anything else is discarded
AI can add new cuisines	Pick from the list, or add an entry that is missing

Under both strategies the AI's answers are matched against the existing list first, so a slight misspelling lands on the entry that already exists rather than creating a near-duplicate. The list itself is managed under Settings => Admin => AI & Processing => Cuisines; see Recipe provenance.

Run it on your whole library

Automatic enrichment only runs when a recipe is created, so recipes imported before you enabled a switch, or before an enrichment kind existed, never catch up on their own. Settings => Admin => AI & Processing => Bulk Enrichment => Enrich All Recipes closes that gap: it queues every enrichment kind whose automatic switch is enabled, for every recipe on the server, under the same rules as the automatic run, supplied data wins and only gaps are filled.

The action asks for confirmation first, because it can be an expensive operation: with many recipes it may take a long time and, on a paid AI provider, use a significant amount of credits. It replaces the old Categorize All Recipes button, which ran only categorization and ignored the switches.

Turning it all off

AI_ENABLED=false (or the global switch in the admin settings) suppresses every enrichment, automatic and manual. No AI request can bypass it.

Prompts

Every AI feature runs from an administrator-editable prompt, nine in total, listed together under Settings => Admin => AI & Processing => Prompts: recipe extraction, image extraction, unit conversion, nutrition estimation, auto-tagging, auto-categorization, allergy detection, Recipe Provenance, and Ingredient Linking. What you see there is exactly what is tunable; there are no hardcoded prompts behind it.

Each feature appends its own input, the recipe under analysis, your household's allergens, the webpage text, after your prompt rather than filling placeholders inside it, so a customised prompt keeps working across upgrades and editing one prompt never changes what a different feature sends. A prompt left empty falls back to the shipped default, and Restore defaults brings all nine back at once.

Video import

Video import downloads the clip with yt-dlp, transcribes the audio, and uses the AI provider to extract the recipe. It requires AI to be enabled: a video import is refused immediately when AI is off, before anything is downloaded or a transcription is billed.

Links from YouTube, Instagram, TikTok, Facebook, Pinterest (including pin.it share links), X, Threads, Snapchat, Vimeo, Dailymotion, Douyin, Bilibili, and RedNote are recognised as videos and take this pipeline; a link from any other site imports as a regular webpage.

Variable	Description	Default
VIDEO_PARSING_ENABLED	Enable the video parsing pipeline	false
VIDEO_MAX_LENGTH_SECONDS	Maximum accepted video length	120
YT_DLP_VERSION	yt-dlp release a development install downloads on first use (the Docker image ships its own)	2026.07.04
YT_DLP_BIN_DIR	Folder containing the yt-dlp binary	./.runtime/bin (dev), /app/bin (prod)
YT_DLP_PROXY	HTTP/SOCKS proxy URL for yt-dlp downloads	(empty)

Photo posts and reels

An Instagram or Facebook post with no video is imported from its caption alone, which only works when the caption holds the whole recipe. Norish decides which path to take by asking yt-dlp whether the post has a video stream.

A post yt-dlp says nothing about either way is treated as a video and downloaded; only a post it reports as having no video, or one where there turned out to be nothing to download, falls back to the caption. Silence is never read as "no video".

If reels still import as photo posts on your instance, check which yt-dlp you are running first: a build too old for Instagram's current markup can fail to report the video at all. Settings => Admin => AI & Processing => Video Processing shows the release the server is actually running, it asks the binary, so it is the truth rather than a stored setting, and it is read-only for the same reason. The Docker image ships the binary named above and upgrading Norish upgrades it; a development install downloads whatever YT_DLP_VERSION names, once, the first time it needs it.

If that field reports no yt-dlp binary found, there is nothing to import with. In Docker, check that YT_DLP_BIN_DIR points at the image's own /app/bin, an empty volume mounted over it hides the shipped binary. On a development install, run an import once with network access and Norish downloads the binary itself; if that fails, place the release named by YT_DLP_VERSION in YT_DLP_BIN_DIR by hand and make it executable.

Transcription

Transcription turns the video's audio into text for the AI step.

Variable	Description	Default
TRANSCRIPTION_PROVIDER	Transcription provider	disabled
TRANSCRIPTION_ENDPOINT	Transcription endpoint (local/custom providers)	(empty)
TRANSCRIPTION_API_KEY	Transcription API key	(empty)
TRANSCRIPTION_MODEL	Transcription model	whisper-1

When the endpoint or API key is left empty, transcription falls back to the AI configuration's endpoint and key, and it follows AI_TIMEOUT_MS the same way. There is no separate transcription timeout: the one number you tuned for your model applies here too, so a hung transcription endpoint gives up instead of holding a video import worker until the server is restarted.