Try every Passio route on your own food, with real token costs.Get a free demo account →

Passio
Start building
Changelog

What's new in Passio.

Release notes for the Nutrition AI Platform, as shipped to production. Account holders see the same notes under What's New in the accounts portal.

Read the docs

October 2026

  1. Catalog Manager: Improved review workflows for edits, labels, and images

    We've updated the Catalog Manager to provide clearer guidance when editing products that require Passio review. The interface now explicitly indicates when changes will be submitted for review, stages label additions to be approved alongside product edits, and clarifies permissions around replacing existing images.

    • Clearer "Submit for review" actions when editing products you didn't add.
    • Labels added during a reviewed edit are now submitted with the change request.
    • Image controls now explain when existing photos or brand logos cannot be replaced.

  2. Catalog Manager opened to all Platform tiers and Easy AI chain routes documented

    The Catalog Manager is now available to all Nutrition Platform subscriptions, allowing more users to research and add products using AI. The research and review experience has been significantly improved with upfront cost estimates, clearer feedback on skipped inputs, and a streamlined review flow. Additionally, the asynchronous Easy AI chain execution endpoints are now officially documented in the OpenAPI specification.

    • Catalog Manager is now available to all Nutrition Platform tiers.
    • Added upfront token cost estimates before running AI research in the Catalog Manager.
    • Improved the Catalog Manager review flow with automatic duplicate merging and clearer status groups.
    • Edits to existing catalog products added by others are now submitted for Passio review.
    • Documented the asynchronous Easy AI chain execution endpoints in the OpenAPI spec.

  3. Catalog Manager workflow upgrades and Easy AI chain API docs

    This release introduces major workflow improvements to the Catalog Manager, including inline cost estimates, better handling of skipped inputs, and a streamlined review process. We've also published official OpenAPI documentation for executing and monitoring asynchronous Easy AI chains.

    • Official API documentation for Easy AI chain execution and status polling.
    • Catalog Manager now shows inline token cost estimates before running AI research.
    • Streamlined Catalog Manager review flow automatically combines duplicates and closes unchanged merges.
    • Edits to shared catalog products are now routed through a Passio review process to ensure data quality.

  4. Kitchen and Health AI Tools: Recipes, Menu Coach, and Multimodal Inputs

    We've introduced seven new AI-powered kitchen and health tools, including pantry-to-recipe generation, recipe makeovers, menu coaching, and daily nutrition reviews. These tools support multimodal inputs—including text, audio, video, and up to four images per request—and automatically link generated recipe ingredients to Passio's food database for accurate nutrition estimates.

    • Seven new AI tools for recipes, menu coaching, and health condition checks.
    • Multimodal support: send text, audio, video, or up to 4 images per request.
    • Generated recipes automatically link ingredients to Passio foods for per-serving nutrition.
    • Menu Coach analyzes restaurant menus to recommend items fitting user goals.

  5. Interactive Demo Guide and AI Studio Rename

    This release introduces a new interactive demo and API guide accessible directly from the dashboard, automatically signing you in with your active API key. We've also renamed 'Easy AI' to 'AI Studio' across the platform to better reflect its comprehensive toolset.

    • Launch the interactive demo and API guide directly from your Account Info tab.
    • 'Easy AI' is now 'AI Studio' across the dashboard and subscription pages.

  6. API documentation stability and internal analytics enhancements

    This release resolves a potential loading issue with the API documentation to ensure reliable access to the API reference. We have also made significant enhancements to our internal analytics infrastructure to improve usage tracking and billing projections.

    • Fixed a potential loading issue with the API documentation.
    • Enhanced internal analytics infrastructure for usage and billing projections.

  7. Usage tracking accuracy and auto-refill improvements

    This release improves the accuracy of usage tracking for concurrent API calls and refines auto-refill behavior to prevent excessive charge attempts when a payment method is declined. It also includes fixes for billing period insights and enhancements to refill projections.

    • Fixed an issue where concurrent API calls could be undercounted in usage totals.
    • Improved auto-refill logic to pause attempts after a declined payment, preventing excessive failed charges.
    • Adjusted the auto-refill trigger for small budgets to ensure sufficient headroom for large requests.
    • Fixed a bug in Route Insights that could include previous cycle usage in the current billing period.
    • Improved the accuracy of projected cycle refills in Route Insights.

  8. Feedback API, Request Tracing, and New AI Models

    This release introduces a new feedback API that lets you report user outcomes on AI results to drive continuous improvement. We've also significantly expanded our model catalog with the latest releases from OpenAI, Anthropic, Google, and xAI, alongside new cost- and speed-optimized model presets.

    • New POST /napi/feedback endpoint to report user actions on AI results.
    • Every API response now includes an X-Request-ID header for tracing and feedback.
    • Added support for GPT-6, Claude 5.5, Gemini 3.8 Flash, and Grok 4.7.
    • Introduced model presets like gpt-6-sol-fast for optimized workloads.

September 2026

  1. EasyAI Studio integration and redesigned Usage Breakdown

    We've integrated EasyAI Studio directly into the Nutrition Platform dashboard, offering an AI assistant to help you build and troubleshoot REST API and Mobile SDK integrations. This release also introduces a completely redesigned Usage Breakdown section for deeper insights into your API consumption, alongside usability improvements to the API Keys interface.

    • New EasyAI Studio assistant to scaffold and troubleshoot API and SDK integrations.
    • Redesigned Usage Breakdown with cost-by-route analysis and top consumer stats.
    • Improved API Keys interface with clearer copy actions and expiration dates.

  2. Real-time token meter and auto-refill billing fixes

    This release introduces a real-time token usage meter in the portal, providing immediate visibility into your API consumption. We have also deployed significant improvements to our auto-refill billing logic, ensuring accurate invoicing and uninterrupted access during sudden usage spikes.

    • Token usage meter now updates in real-time without a 15-minute delay.
    • Visual scaling of the usage meter improved to prevent projections from compressing current usage.
    • Fixed an issue where auto-refills could trigger duplicate invoices during concurrent requests.
    • Improved payment recognition to prevent access interruptions while waiting for provider confirmations.

  3. Affiliate Attribution Engine Upgrade & Admin Reconcile Safety

    We've significantly upgraded the affiliate link attribution engine to automatically capture and credit more conversions, especially those from delayed purchases or direct URLs. For Passio admins, we've introduced robust safety guardrails and UI enhancements for manual reconciliations, alongside an automated backlog retry system to clear unmatched actions faster.

    • Upgraded affiliate attribution engine to automatically credit delayed or direct-URL conversions.
    • Introduced safety guardrails for manual unmatched action reconciliation to prevent misattribution.
    • Added an automated backlog retry pass to continuously re-evaluate unmatched actions.
    • Enhanced the Admin Affiliate Earnings UI with deeper diagnostic data and streamlined workflows.

  4. Improved Japanese localization and portion sorting

    We've significantly improved the quality of Japanese localization by fixing search translation mismatches, preserving brand names, and prioritizing metric portion sizes. Under the hood, we've also upgraded the AI translation model to improve terminology discipline.

    • Improved Japanese search accuracy by fixing translation mismatches for common foods.
    • Preserved original brand names in Japanese results instead of transliterating them.
    • Reordered portion sizes in Japanese responses to prioritize metric units over US customary units.

  5. OpenAPI Intent Extensions & Usage Caching Improvements

    This release significantly enhances the OpenAPI specifications with new intent-mapping extensions, expanded schema descriptions, and a comprehensive endpoint routing guide. Under the hood, we've overhauled usage tracking and caching mechanisms to prevent cache stampedes and reduce BigQuery costs.

    • Added `x-passio-intents` and `x-passio-not-for` extensions to OpenAPI specs for better AI intent mapping.
    • Added a comprehensive "Which endpoint should I call?" guide to the API documentation.
    • Affiliate link conversion endpoints now accept full meal plan objects.
    • Implemented atomic Redis counters and singleflight request collapsing to prevent BigQuery cost spikes.

August 2026

  1. Affiliate earnings timeseries, Route Insights search, and Junk Food Meter supplement handling

    This release introduces a new timeseries view for affiliate earnings, providing daily insights and projections for links, clicks, and conversions. We've also added search and filtering capabilities to the Route Insights dashboard and updated the Junk Food Meter to correctly identify and handle dietary supplements.

    • View daily affiliate earnings timeseries, conversion funnels, and 30-day projections.
    • Search and filter API routes in the Route Insights dashboard.
    • The Junk Food Meter detects dietary supplements and returns a specific disclaimer instead of a processing score.

  2. Simplified Grocery Replacement Configuration

    We streamlined how grocery replacements are tested in the Catalog Manager. The testing tools automatically sync with the centralized platform configuration, ensuring test results perfectly match the SDK behavior without requiring manual setup.

    • Removed manual chain ID configuration for grocery replacements in the Catalog Manager.
    • The Grocery Replacement test tab automatically mirrors the SDK's centralized configuration.

  3. Image Enrichment, AI Context Messages, and Performance SLOs

    This release introduces optional product and brand image enrichment across our food-lookup APIs and adds support for caller-supplied context messages in all AI requests. We've also published a comprehensive API performance and SLO guide, and added a detailed integration guide for the grocery replacement chain.

    • Enrich API responses with product and brand images using the new X-Passio-Product-Images header.
    • Provide additional context to AI models using the optional message field for multimedia and text requests.
    • Access our new API Performance and SLO reference guide for configuring client timeouts and alerts.
    • View Junk Food Meter (jfm_score) scores in grocery replacement chain results.

July 2026

  1. Enhanced server-side resolution for grocery catalog features

    We've updated the Passio Nutrition AI Platform to improve how grocery replacements and catalog items are processed. These backend enhancements ensure more reliable catalog ingestion, merging, and alternative item finding.

    • Improved server-side resolution for grocery replacements.
    • Enhanced reliability for catalog ingestion and merging.

  2. Fix for "My Products" visibility and improved regional semantic search

    This release fixes a bug that prevented custom products from appearing in the "My Products" list and web link searches. It also enhances the reliability of regional semantic search with better fallback handling and search term processing.

    • Fixed an issue preventing custom catalog products from appearing in "My Products".
    • Improved reliability and accuracy of regional semantic search.

  3. Food Tags API, Tag Search Filters, and Relevance Improvements

    This release introduces a new Food Tags API and powerful tag-based filtering for search results, allowing you to precisely include or exclude foods based on diet, allergens, and claims. We've also deployed significant search relevance improvements to better surface verbose branded names and automatically boost products matching dietary intents like 'keto' or 'vegan'.

    • New Food Tags API to retrieve the canonical taxonomy of food tags.
    • Added positive and negative tag filtering to the Passio-Search-Config header.
    • Improved search relevance for long branded names and dietary-intent queries.

June 2026

  1. Improve Edamam search result quality by filtering placeholder entries

    We've improved the quality of food and recipe search results from the Edamam integration. The API now automatically filters out entries with generic placeholder names, ensuring cleaner and more relevant data for your users.

    • Filtered out placeholder food and recipe names from Edamam search results.
    • Improved overall search relevance and data cleanliness.

  2. Add standardized country codes to enriched search results

    Enriched search results now include a standardized ISO2 country code for products. This new field is derived from raw brand data and provides consistent regional context for AI functions and market filtering.

    • New `country` field in enriched search result payloads.
    • Standardized ISO2 format (e.g., 'US', 'CA') for reliable regional filtering.
    • Prioritized resolution for US and Canadian markets.

  3. Advanced Search Enrichment and Visual Extraction Improvements

    This release introduces an opt-in enrichment mode for advanced product searches, providing AI-ready nutritional scoring and ingredient data. It also improves the visual food extraction tool to accurately identify and reject non-food items, and includes visual and formatting enhancements for branded transactional emails.

    • Added an `enrich` option to advanced search for AI-tailored ingredient and nutrient scoring.
    • Visual food extraction now explicitly detects non-food items and returns a descriptive error.
    • Optimized plain-text formatting and updated background colors for branded transactional emails.

  4. Application Branding Refresh and Email Deliverability Improvements

    The Passio Accounts portal and email templates feature a refreshed green color palette and new logo assets. Under the hood, significant improvements were made to email deliverability and compliance, alongside enhancements to internal admin tools.

    • Refreshed application UI and email templates with a new green brand color palette.
    • Improved email deliverability and compliance for automated communications.

May 2026

  1. Improved usage projections, no-cancel subscription workflow, and API docs enhancements

    This release introduces a more forgiving subscription failure workflow that preserves licenses during payment issues, alongside more accurate V2 usage projections in the dashboard. We've also significantly enhanced our API documentation with clearer guidance on modern endpoints and legacy tools, and improved how usage notifications behave after capacity refills.

    • Subscriptions are no longer canceled upon payment failure; token issuance is now temporarily suspended until the invoice is paid.
    • V2 usage projections in the dashboard are now calculated using an "effective days" window for greater accuracy.
    • Capacity usage notifications (80% and 100%) now automatically re-arm after a subscription refill.
    • API documentation has been comprehensively updated with better descriptions and a new "Legacy AI Tools" category.

  2. Affiliate Earnings Dashboard, New AI Models, and Billing CC Management

    This release introduces a comprehensive Affiliate Earnings dashboard and a new API for generating tracked Instacart shopping links. We've also added support for the latest AI models including GPT-5.5 and Claude Opus 4.7, increased the Mobile AI free user limit, and added the ability to manage billing CC email recipients.

    • Generate tracked Instacart affiliate links and drop-in CTA buttons.
    • Track affiliate conversions and manage payouts via Stripe Connect.
    • Access new flagship AI models: GPT-5.5, Claude Opus 4.7, Grok 4.3, and Gemini 3.1 Flash Lite.
    • Manage billing CC email addresses directly from the Team page.
    • Mobile AI free user limit increased from 3 to 10.

April 2026

  1. Enhanced LLM routing resilience, smart model fallbacks, and accurate billing tracking

    This release introduces a major upgrade to our AI model routing, adding intelligent fallback mechanisms that automatically switch to alternative models if a primary model experiences rate limits or timeouts. We've also improved system resilience against external API disruptions and updated our API responses to accurately report the specific model used for each request, ensuring transparent billing.

    • Intelligent model fallbacks automatically reroute requests if an AI provider fails or times out.
    • API responses now include the exact model used (`model_used`) for accurate billing transparency.
    • Improved license validation caching ensures continued service during temporary external API outages.
    • Optimized timeout handling dynamically adjusts wait times based on the complexity of the requested AI model.

  2. Expanded AI Model Selection and Enhanced Portal Experience

    This release introduces a massive expansion of our supported AI models, bringing the latest generations from OpenAI, Google, Anthropic, and xAI to the platform. We've also revamped the model selection interface in the portal to make it easier to compare and choose the right model for your needs. Under the hood, significant upgrades to our LLM infrastructure improve response reliability and parsing accuracy.

    • Added support for the latest AI models, including GPT-5.4, Gemini 3 & 3.1, Claude 4.5 & 4.6, and Grok 4.
    • Revamped the portal's model selection interface to group by provider and display intuitive cost tiers.
    • Improved AI response reliability by enforcing structured JSON outputs directly at the model level.

  3. OpenAPI specification, multi-source barcode lookup, and infrastructure optimizations

    This release introduces a comprehensive OpenAPI 3.0.3 specification and detailed documentation for API headers and billing transparency. We've enhanced barcode lookups to cascade across multiple data sources and improved the resilience of visual extraction and semantic search. Internally, we've optimized database connection pooling, LLM client timeouts, and Redis caching infrastructure for better performance and stability.

    • Comprehensive OpenAPI 3.0.3 specification and API header guides.
    • Multi-source barcode lookup cascading through Passio and Edamam.
    • Improved visual extraction handling for non-food items.
    • SDK Assistant chatbot added to the platform portal.
    • Optimized Redis caching with LFU eviction and removed unused replicas.