lfnovo/open-notebook
47.3
Weak · 21 October 2026
56.2k
lines of production code
TypeScript
with Python
3
measurements over time
What this system is
Open Notebook is a self-hosted, AI-powered knowledge management platform that enables users to organize information into notebooks, ingest diverse sources, and interact with their data through chat and search. The system leverages a modular backend to orchestrate RAG workflows, podcast generation, and document transformations, while a modern Next.js frontend provides a responsive interface for managing credentials, models, and content. It supports local and cloud deployments with optional AI extraction engines, offering a secure, containerized environment for private data analysis.
How it got here
2024–2025 — Full-stack architecture modernization
47 changes.
The project underwent a comprehensive overhaul, migrating the backend from a monolithic Streamlit structure to a modular FastAPI application with robust security and async database handling. The frontend was rebuilt using Next.js 16 with a new design system, replacing legacy UI components with a structured dashboard and reusable libraries. This period focused on stabilizing the core infrastructure, enhancing security, and establishing a clean, maintainable codebase for future feature development.
2026 — AI infrastructure and security hardening
13 changes.
This period focused on centralizing AI provider management through a database-backed credential system and model registry, replacing scattered configuration methods. Significant security improvements were implemented, including fixing Jinja2 template injection vulnerabilities, enforcing strict path handling for podcasts, and securing local development environments. The work also enhanced the user experience with real-time SSE streaming, expanded internationalization support, and introduced comprehensive testing and deployment tooling.
Features
Dashboard layout with authentication guard and command palette
The dashboard route now uses a layout component that enforces authentication, redirecting unauthenticated users to the login page while preserving the intended destination in session storage. During the initial auth check, a loading spinner is displayed. The layout also wraps the dashboard content with an error boundary, modal provider, and create-dialogs provider, and includes a new CommandPalette component for quick navigation and search. The main dashboard page itself now immediately redirects users to the notebooks section.
frontend/src/app/(dashboard) · high confidence
Expanded language support and i18n validation
The application now offers a broader selection of user interface languages, adding Bengali (bn-IN), Catalan (ca-ES), German (de-DE), and Turkish (tr-TR) to the existing set of supported locales. To ensure translation quality and consistency, the frontend now includes automated tests that verify key parity, interpolation placeholder alignment, and the absence of unused translation keys across all languages.
frontend/src/lib/locales · high confidence
Introduce Surreal-commands worker integration for background processing
The application now routes embedding, source processing, and podcast generation tasks through a dedicated background worker using the \surreal-commands\ library. This change introduces new command modules (\embedding\_commands\, \source\_commands\, \podcast\_commands\) that handle asynchronous execution, including robust retry strategies for transient errors and permanent failure handling for deleted or invalid records. Additionally, the worker initialization ensures internal SurrealDB websocket traffic bypasses HTTP proxies to prevent connection issues.
commands · high confidence
Introduce dedicated system prompt for Source Chat
A new system prompt file (system.jinja) has been added for the Source Chat feature. This prompt configures the AI to act as a specialized research assistant focused on analyzing a specific source document. It instructs the model to leverage both the source content and any generated insights, enforce specific citation formats using document and insight IDs, and use $$ delimiters for math rendering. The prompt also includes instructions to handle cases where source text is unavailable and to maintain focus on the specific document context.
_prompts/source\chat · high confidence
Introduce new API router structure and capabilities endpoint
The API now exposes a dedicated /capabilities endpoint that reports the runtime availability of opt-in extraction engines (Docling and Crawl4AI), allowing the frontend to dynamically gate engine options and the OCR toggle based on what is actually importable in the container. Additionally, the API router structure has been reorganized into a modular layout with separate files for authentication, credentials, chat, commands, and other domains, improving maintainability and separation of concerns.
api/routers · high confidence
New Advanced settings page with embedding rebuild and system info
A new Advanced page has been added to the dashboard, providing two key capabilities: a System Info panel that displays the current application version, checks for available updates, and links to the GitHub repository, and a Rebuild Embeddings tool that allows users to re-index their data. The rebuild feature supports selecting specific content types (sources, notes, insights) and offers two modes (existing or all), with real-time progress tracking and status polling during the operation.
frontend/src/app/(dashboard)/advanced · high confidence
New Docker Compose examples for local AI and EasyPanel deployment
The examples directory now provides ready-to-use Docker Compose configurations for various deployment scenarios. Users can choose between a fully local setup with Ollama and Speaches for privacy-focused AI, a combined local speech processing setup, or a single-container mode for constrained environments. Additionally, an EasyPanel template is included to simplify one-click deployment of Open Notebook with SurrealDB, handling secret generation and service wiring automatically. The development setup has also been updated to bind SurrealDB to localhost only for security and align ports with the standard production configuration.
examples · high confidence
New Podcasts dashboard page with Episodes and Templates tabs
A new Podcasts page has been added to the dashboard, providing a unified view for managing podcast content. The page features two tabs: 'Episodes' and 'Templates', allowing users to switch between viewing episode profiles and managing podcast templates. An alert is displayed if any episode or speaker profiles require model setup, guiding users to complete necessary configurations before proceeding.
frontend/src/app/(dashboard)/podcasts · high confidence
New Settings page added to the dashboard
A new Settings page has been introduced at the dashboard settings route. It displays a title and a refresh button, and renders a SettingsForm component to allow users to manage their settings.
frontend/src/app/(dashboard)/settings · high confidence
New Sources management page with sorting and infinite scroll
The Sources page has been implemented to allow users to view, sort, and manage their data sources. Users can now sort the list by any column and navigate through large datasets using infinite scroll, which automatically loads more items as they reach the bottom of the list. The page also includes keyboard navigation for accessibility and a dedicated dialog for adding new sources.
frontend/src/app/(dashboard)/sources · high confidence
New UI component library and secure Markdown editor
The frontend now includes a comprehensive set of new UI components (Accordion, AlertDialog, Alert, Badge, Button, Card, Checkbox, Collapsible, Command, Dialog, DropdownMenu, FormSection, Input, Label) built on Radix UI primitives, providing a consistent design foundation. Additionally, a new MarkdownEditor and MarkdownRenderer are introduced, featuring syntax highlighting for over 20 programming languages, LaTeX math rendering via KaTeX, and strict sanitization of raw HTML in previews to prevent script injection while preserving safe content like code blocks and math.
frontend/src/components/ui · high confidence
New common UI components and language switching UX
The \frontend/src/components/common\ directory now includes a suite of reusable UI components: \CommandPalette\ for quick navigation and search via keyboard shortcuts, \LanguageToggle\ and \LanguageLoadingOverlay\ to manage language switching with a loading indicator, \ThemeToggle\ for light/dark/system theme selection, \ConfirmDialog\ for safe destructive actions, \ConnectionGuard\ to handle API connectivity errors, \ContentUnavailable\ for consistent empty/error states, \ContextIndicator\ and \ContextToggle\ to manage source/note context modes, \InlineEdit\ for in-place editing, \ModelSelector\ for AI model selection, and basic utilities like \LoadingSpinner\, \EmptyState\, \ErrorBoundary\, and \LoginForm\. These components standardize the user interface for authentication, settings, and content interaction across the application.
frontend/src/components/common · high confidence
New credential-based API key management in Settings
The Settings page now uses a credential-based system for managing provider configurations. This introduces a new UI with a ProviderSection that lists providers and their credentials, allowing users to add, edit, test, and delete API keys via CredentialFormDialog and CredentialItem components. It includes features like model discovery (DiscoverModelsDialog), credential migration from environment variables (MigrationBanner), and safe deletion with model migration options (DeleteCredentialDialog). Default model selection is also refined with DefaultModelSelectors, supporting optional slots with fallback hints and embedding model change warnings (EmbeddingModelChangeDialog).
frontend/src/components/settings · high confidence
New episode and speaker profile configuration forms
Users can now create and edit podcast episode and speaker profiles through dedicated dialog forms. The episode profile form allows configuration of LLM models for outlining and transcription, language selection, segment counts, and optional token limits, while enforcing validation rules such as minimum segment counts. The speaker profile form enables management of voice models and individual speaker configurations, including name, voice ID, backstory, and personality traits, with support for up to four speakers per profile. Both forms integrate with the existing podcast API for creating and updating profiles.
frontend/src/components/podcasts/forms · high confidence
New frontend hooks for authentication, credentials, models, and UI state management
The \frontend/src/lib/hooks\ directory now includes a comprehensive set of new React hooks to manage application state and API interactions. These include \use-auth\ for handling login/logout and hydration-based auth checks, \use-credentials\ for managing API keys and provider configurations, and \use-models\ for model registry operations. UI state is now centralized via \use-modal-manager\ for URL-based modal routing and \use-create-dialogs\ for context-driven dialog management. Additional hooks like \use-providers\, \use-capabilities\, \use-search\, \use-notebooks\, \use-notes\, \use-podcasts\, and \use-source-chat\ provide structured access to their respective backend APIs, while \use-ask\ handles the core AI query streaming logic with idle-timeout safeguards. New utility hooks \use-media-query\ and \use-navigation\ support responsive design and return-path tracking.
frontend/src/lib/hooks · high confidence
New notebook creation dialog and collapsible column component
Added a new CreateNotebookDialog component that allows users to create notebooks via a form with name and description fields, integrated with validation and translation support. Also introduced a CollapsibleColumn component that provides a toggleable sidebar interface with proper text orientation handling for CJK characters.
frontend/src/components/notebooks · high confidence
New open\_notebook.utils package for shared infrastructure
The open\_notebook/utils package introduces a centralized set of utilities for the application. It includes a ContextBuilder for assembling and managing context from sources, notes, and insights with token-aware truncation and deduplication. Chunking and embedding modules provide content-type-aware text splitting and batched embedding generation with configurable batch sizes and retry logic. Encryption utilities now use versioned PBKDF2 key derivation for secure API key storage while maintaining backward compatibility. Additional modules handle LLM error classification, proxy configuration to bypass internal database connections, runtime capability detection for optional engines like Docling and Crawl4AI, SSL configuration for connection tests, and text processing for handling AI thinking tags.
_open\notebook/utils · high confidence
New release-candidate verification tooling for local image testing
Added executable scripts and configuration files in the release-test directory to validate published Docker images locally. The rc-stack.sh script allows developers to spin up a release candidate image with an optional data dump and optional heavy runtimes (Docling, Crawl4AI) for manual verification, while release-image-test.sh automates fresh-install and upgrade scenarios to ensure database migrations and worker processes function correctly in the shipped container. These tools use a dedicated docker-compose setup with an nginx reverse proxy to isolate release testing from the development environment.
scripts/release-test · high confidence
New search UI components: model configuration, notebook scoping, and streaming results
This change introduces several new components in the search interface. The AdvancedModelsDialog allows users to configure specific models for strategy, answer, and final answer generation. The NotebookScopeSelector enables filtering search results by selecting specific notebooks. The SaveToNotebooksDialog provides a way to save search results to selected notebooks. The StreamingResponse component renders the search process with collapsible sections for strategy reasoning, individual answers, and the final answer, including clickable reference links.
frontend/src/components/search · high confidence
New source creation steps for notebooks, processing, and source type selection
Added three new components to the source creation workflow: NotebooksStep allows users to select existing notebooks to associate with a source; ProcessingStep provides controls for applying transformations and configuring embedding settings based on system configuration; and SourceTypeStep enables users to choose between link, upload, or text input methods, including support for batch URL/file uploads and HTML content detection in text fields.
frontend/src/components/sources/steps · high confidence
New source detail page with integrated chat
A new page component has been added at frontend/src/app/(dashboard)/sources/\[id\]/page.tsx to display individual source details. This page implements a two-column layout: the left side renders the SourceDetailContent component, while the right side integrates a ChatPanel for interactive conversations. The page utilizes the useSourceChat hook to manage chat state, including message streaming, session management, and model overrides, and includes navigation logic to return to the previous path via a back button.
frontend/src/app/(dashboard)/sources/\[id\] · high confidence
New source management components and chat interface
This change introduces a new set of components for managing sources and interacting with them via chat. Users can now add existing sources to notebooks using the AddExistingSourceDialog, which deduplicates search results by parent ID. The AddSourceDialog provides a wizard-based interface for creating new sources (links, uploads, text) with validation and async processing support. The ChatPanel component enables conversational interaction with sources, including session management, model selection, and message actions like copying or saving to notes. SourceCard displays source status and actions, while NotebookAssociations allows linking sources to multiple notebooks.
frontend/src/components/sources · high confidence
New utility scripts and Docker entrypoint for documentation, link checking, and optional runtimes
This change introduces several new scripts to the \scripts/\ directory. \export\_docs.py\ consolidates markdown documentation files from subdirectories into single files with a Table of Contents, while \check\_md\_links.py\ validates relative links in tracked markdown files. A new \docker-entrypoint.sh\ enables opt-in installation of heavy extraction runtimes (Docling and Crawl4AI) at container startup via environment variables, handling dependencies and caching. Additionally, \wait-for-api.sh\ ensures the API is healthy before the frontend starts, and a \README.md\ documents these new scripts.
scripts · high confidence
Notebook page introduces bulk source/note context controls and mobile tabbed navigation
The notebook detail page now allows users to bulk include or exclude all sources and notes from the chat context, with the selected action persisting as the default for any additional items loaded later. On mobile devices, the layout switches to a tabbed interface (Sources, Notes, Chat) to avoid double-mounting components, while desktop retains the multi-column view.
frontend/src/app/(dashboard)/notebooks/\[id\] · high confidence
Notebooks UI components and tests added
This change introduces the frontend components for the notebooks dashboard, including the list view (NotebookCard, NotebookRow, NotebookList), the notebook header with inline editing and archive/delete actions (NotebookHeader, NotebookDeleteDialog), and the notebook detail columns for notes and chat (NotesColumn, ChatColumn). It also adds the NoteEditorDialog for creating and editing notes, along with a RecentlyViewed section for quick access to recent items. Comprehensive unit tests are included for ChatColumn, NoteEditorDialog, NotebookHeader, and NotesColumn to verify loading states, error handling, context mode toggling, and data persistence.
frontend/src/app/(dashboard)/notebooks/components · high confidence
Notebooks page with list/tile view toggle and recent items
The notebooks dashboard now displays a 'Recently Viewed' section at the top, followed by active and archived notebook lists. Users can switch between tile and list display modes using the new view toggle buttons, search notebooks by name, and create new notebooks via a dialog. The page also includes a refresh button to reload the notebook data.
frontend/src/app/(dashboard)/notebooks · high confidence
Removals
Removal of legacy Streamlit UI components
The \stream\_app\ module has removed several core files (\chat.py\, \consts.py\, \note.py\, \source.py\, \utils.py\) that previously provided the chat interface, note and source management dialogs, context configuration, and session state setup. This deletion eliminates the direct Streamlit-based UI for interacting with notes, sources, and chat threads within this specific application area.
_stream\app · high confidence
Removal of legacy Streamlit notebook and search pages
The dedicated Streamlit-based pages for managing notebooks (2\_📒\Notebooks.py) and searching the knowledge base (3\🔍\_Search.py) have been removed from the application. This eliminates the previous UI components that allowed users to create, edit, and delete notebooks, as well as perform text and vector searches across sources and notes, indicating a shift away from the standalone Streamlit interface for these features.
pages · high confidence
Security
Fix: Prevent Jinja2 template compilation of user-controlled prompts
A new generic Jinja2 pattern file has been introduced to safely render prompts. This change addresses a security vulnerability where user-controlled input was previously compiled as Jinja2 template source, potentially allowing injection attacks. The new pattern ensures that the \prompt\ variable is rendered as literal text rather than being interpreted as template code, while still conditionally appending \format\_instructions\ if present.
prompts/pattern · high confidence
Secure runtime API URL configuration with strict header validation
The frontend now exposes a runtime configuration endpoint that determines the API base URL by validating the incoming Host and X-Forwarded-Proto headers against strict hostname and protocol patterns. This change prevents header-injection attacks where a malicious proxy could redirect browser API traffic (including auth tokens) to an attacker-controlled host. The endpoint prioritizes an explicit API\_URL environment variable, falls back to auto-detection from validated request headers, and safely defaults to localhost if validation fails or headers are missing.
frontend/src/app/config · high confidence
Behavioural changes
API server restructured into a standalone FastAPI application with hardened security and configuration
The backend has been refactored from a Streamlit-based structure into a dedicated FastAPI application located in the \api/\ directory. This change introduces a new \PasswordAuthMiddleware\ that enforces Bearer-token authentication using constant-time comparison, while allowing unauthenticated access to health and documentation endpoints. Security is further hardened by adding a \MaxBodySizeMiddleware\ to reject oversized request bodies (defaulting to 100MB) before they are processed, and by ensuring CORS headers are correctly applied to error responses. The application now supports configurable CORS origins via the \CORS\_ORIGINS\ environment variable and includes resilient startup logic that retries database connectivity. Additionally, the codebase now includes a \CommandService\ for managing background jobs and a \PodcastService\ that resolves speaker profiles by record ID, alongside new Pydantic models for notebook scoping in search and ask operations.
api · high confidence
Ask, Chat, and Source workflows rewritten for reliability and security
The Ask, Chat, and Source processing graphs have been completely rewritten to improve reliability, security, and user control. The Ask workflow now uses a three-stage strategy (strategy, search, synthesis) with an increased output budget of 8192 tokens to prevent silent truncation, and fails fast when models return empty or unusable responses. Chat and Source Chat now support model overrides, handle extended thinking models by stripping thinking tags, and include source content in the chat context. The Source graph now uses content-core 2.x with better error handling, including fallbacks for missing engines and improved YouTube transcript language support. The old Ask content graph and related tools have been removed, and user-controlled prompts are now safely rendered as variables rather than compiled as Jinja2 templates, addressing a security vulnerability.
_open\notebook/graphs · high confidence
Async database migration system with SurrealQL injection protections
The database module now uses an async migration runner (async\_migrate.py) that executes SurrealQL migration files (1–25) via the official Python SurrealDB client, with a synchronous wrapper (migrate.py) for backward compatibility. Repository functions (repository.py) now validate table and relationship identifiers against a strict regex to prevent SurrealQL injection in RELATE, UPSERT, and UPDATE operations, and ensure the internal SurrealDB websocket is excluded from HTTP proxies. Database connection details are read from SURREAL\_URL with fallbacks to SURREAL\_ADDRESS/SURREAL\_PORT, and passwords from SURREAL\_PASSWORD with fallback to SURREAL\_PASS.
_open\notebook/database · high confidence
Backend architecture refactoring and new error handling
The Open Notebook backend has been restructured to improve maintainability and reliability. The monolithic \domain.py\, \prompter.py\, \repository.py\, and \utils.py\ modules have been removed and replaced with a modular architecture: domain models are now in \open\_notebook/domain/\, prompts are managed via the \ai\_prompter\ library, database access is handled by a refactored \database\ module with explicit migrations, and shared utilities are distributed across \open\_notebook/utils/\. A new \config.py\ centralizes configuration, including a persistent tiktoken cache and a default LLM timeout of 180 seconds. Additionally, new exception types (\UnsupportedTypeException\, \IncompleteGenerationError\, \ContextLengthExceededError\, \NoTranscriptFound\) have been introduced to provide more granular error handling, particularly for background commands which now correctly avoid retrying context-length errors.
_open\notebook · high confidence
Centralized AI provider management with credential-based key provisioning and connection testing
The AI infrastructure module has been restructured to use a single provider registry as the source of truth, replacing scattered configuration dictionaries. API keys are now managed via database-backed Credential records with environment variable fallback, and a new connection tester validates provider connectivity and model availability using minimal API calls. Model discovery automatically fetches and classifies available models from supported providers, while URL validation is re-checked at request time to mitigate DNS-rebinding risks and prevent filesystem information leakage during credential tests.
_open\notebook/ai · high confidence
Database schema and search logic updates
This update introduces a series of database migrations that enhance search performance and flexibility, introduce credential-based API key management, and refine podcast profile configurations. Search functions now support notebook-scoping and resolve insight hits to parent sources, while new indexes improve source listing speed. A dedicated credential table replaces the previous provider configuration singleton, allowing for more granular API key and model settings. Podcast profiles now use model registry references instead of legacy provider/model strings, and audio file paths are stored relative to the podcast folder. Additionally, the system now tracks recently viewed notebooks and sources, exposes Docling enrichment toggles, and allows per-transformation model selection.
_open\notebook/database/migrations · high confidence
Documentation rewritten to match v1.15.0 codebase
The documentation under \docs/\ has been completely rewritten against the current v1.15.0 codebase to remove outdated or imagined features. The install guide now points to a single canonical path using the shipped \docker-compose.yml\, and the user guide reflects the actual product capabilities (e.g., insights as a first-class concept, correct UI labels, and accurate configuration variables). Configuration and troubleshooting sections have been updated to reference the correct environment variables and error messages, and development documentation now accurately describes the project's architecture, testing, and contribution workflows.
(repo-wide) · high confidence
Domain models restructured with new base class and credential management
The domain layer has been refactored to introduce a new \ObjectModel\ base class in \open\_notebook/domain/base.py\, which centralizes database operations and adds validation for \order\_by\ clauses to prevent SurrealQL injection. The \Credential\ model now supports flexible configuration via a \config\ object (persisting fields like \num\_ctx\ for Ollama) and handles API key decryption, while the legacy \ProviderConfig\ singleton is retained alongside the new per-provider \Credential\ records. New domain models include \ContentSettings\ for managing processing engines and OCR toggles, and \Transformation\ for prompt-based transformations.
_open\notebook/domain · high confidence
Frontend restructured with Next.js 16, increased upload limits, and renamed settings page
The frontend has been upgraded to Next.js 16, which includes an increase in the proxy body size limit to 100MB to support larger file uploads. The settings page previously labeled 'API Keys' has been renamed to 'Models' (route \/settings/models\), with a permanent redirect in place for the old \/settings/api-keys\ path. Additionally, the development server now supports custom hostnames via the \NEXT\_ALLOWED\_DEV\_ORIGINS\ environment variable, and the application is configured for standalone output to optimize Docker deployments.
frontend · high confidence
Improved API URL resolution and credential update reliability
The frontend now uses a centralized configuration module that prioritizes runtime API settings from the server over environment variables, resolving connectivity issues for reverse proxy users. Additionally, credential updates are fixed to correctly clear fields by sending explicit null values instead of undefined, ensuring that emptying a field in the UI actually removes it on the backend.
frontend/src/lib · high confidence
Improved chat context management, error clarity, and localization support
This update introduces several improvements to the frontend's utility layer. Users can now bulk include or exclude all sources and notes from chat context, with sensible defaults (insights when available, otherwise full) that persist across pagination loads. Error messages from the backend are now mapped to user-friendly, localized strings, and sources with no text content are handled gracefully. Additionally, date formatting now supports a wider range of languages via date-fns locale mappings, and KaTeX rendering is more robust against Unicode punctuation in math spans.
frontend/src/lib/utils · high confidence
Introduces new Zustand-based state management stores for authentication, navigation, theming, and UI layout
The application now uses dedicated Zustand stores to manage core frontend state, replacing previous implementations. The new auth-store handles login, logout, and session validation with specific error handling for network issues and server connectivity. A theme-store enables users to switch between light, dark, and system themes, persisting the choice to localStorage and applying it to the document root. Navigation and layout preferences are now managed by separate stores: navigation-store tracks return paths with a 1-hour staleness check, notebook-columns-store persists the collapsed state of sources and notes panels, notebook-view-store saves the list/tile view preference, and sidebar-store manages the sidebar's collapsed state. All stores use persistence middleware to retain user preferences across sessions.
frontend/src/lib/stores · high confidence
Introduction of comprehensive TypeScript type definitions for frontend-backend contracts
The frontend now includes a complete set of TypeScript interfaces in \frontend/src/lib/types\ that strictly define the data structures exchanged with the backend API. This covers core domain models (Notebooks, Notes, Sources), configuration and connection states, authentication, and the model registry. It also introduces specific types for new and enhanced features: podcast episode profiles and speaker configurations (referencing profiles by ID), search and 'Ask' requests scoped to specific notebooks, source chat sessions with context indicators, and per-transformation model selection. These definitions ensure type safety across the application and align the frontend with the current Python API schema.
frontend/src/lib/types · high confidence
Introduction of the 'Quiet Green' design system and typography foundation
The application now uses a new 'Quiet Green' design foundation, defined in the global CSS with a specific color palette (fern, sage, gold, teal, etc.), semantic surface/ink variables, and squared border radii (4–6px). This change updates the root layout to load the new typography stack (Instrument Sans, Bricolage Grotesque, Spline Sans Mono) and includes a new \/dev/design\ page that serves as a living styleguide for these tokens. The visual identity shifts to a red-free, nature-inspired palette where color is used for context rather than washing reading surfaces.
frontend/src/app · high confidence
New application shell and sidebar layout with credential migration guidance
The frontend layout has been restructured with a new AppShell component that wraps the main content area, and a comprehensive AppSidebar featuring collapsible navigation, platform-aware keyboard shortcuts, and updated menu items (including the renamed 'Models' page). A new SetupBanner component now appears at the top of the main content, alerting users if encryption is not configured or if they need to migrate provider credentials from environment variables to the new credential-based system.
frontend/src/components/layout · high confidence
New centralized API client and modular service layer
The frontend now uses a unified API client (axios) with a configurable request timeout (defaulting to 10 minutes, adjustable via NEXT\_PUBLIC\_API\_TIMEOUT\_MS) and automatic authentication header injection. This client powers a new modular service layer in frontend/src/lib/api, introducing dedicated modules for capabilities, chat, credentials, embeddings, insights, models, notebooks, notes, podcasts, providers, search, settings, source-chat, sources, and transformations. This change standardizes how the frontend communicates with the backend, replacing previous ad-hoc fetch calls with a consistent, typed, and retry-aware service architecture.
frontend/src/lib/api · high confidence
New podcast generation prompts with solo-speaker mode and language enforcement
The podcast generation workflow now uses dedicated Jinja2 prompt templates for outlining and transcript creation. The outline prompt enforces strict language constraints (generating content in the specified language while keeping JSON keys in English) and supports extended thinking tags for reasoning. The transcript prompt introduces a solo-speaker mode that prevents the model from inventing additional speakers when only one is provided, and it also enforces language consistency for dialogue. Both prompts ensure raw JSON output without code blocks and prevent placeholder content.
prompts/podcast · high confidence
New structured RAG query process with improved citation and math formatting
The Ask feature now uses a new three-stage prompt strategy (entry, query\_process, final\_answer) to improve research accuracy. The system first formulates a search strategy, then generates answers for specific terms, and finally synthesizes a response. This update enforces stricter citation rules, requiring exact document IDs (including type prefixes like 'note:' or 'source:') to be used in brackets, and introduces specific formatting instructions for math (using $$ for display and $ for inline) to ensure proper rendering.
prompts/ask · high confidence
New system prompt for chat assistant with math and citation guidance
A new system prompt file (prompts/chat/system.jinja) has been introduced to define the behavior of the chat assistant. This prompt instructs the model to act as a cognitive study assistant, specifically directing it to use $$...$$ for display math and $...$ for inline math to ensure proper rendering. It also enforces strict citation rules, requiring the assistant to reference documents using their exact IDs (including type prefixes like 'note:' or 'source:') found in the provided context, and prohibits making up document IDs.
prompts/chat · high confidence
Podcast generation and management UI overhaul
The podcast workspace has been rebuilt with a new content selection interface that allows users to choose specific notebooks, sources, and notes for generation, including granular source modes (insights vs. full content) and real-time token/character counting. Episode cards now display resolved model details (provider and model name) for both new and legacy episodes, with graceful fallbacks for unresolvable references. Profile management panels for episodes and speakers have been introduced, enabling users to create, edit, duplicate, and delete profiles, with clear indicators for required model setup and voice model configuration. The generation dialog has been refactored to handle content selection and context building more robustly, addressing previous token-count race conditions.
frontend/src/components/podcasts · high confidence
Podcast module initialization with secure path handling and model registry integration
The new podcasts module introduces strict security controls for audio file storage by enforcing relative paths within the configured podcasts folder, preventing path traversal vulnerabilities during both generation and retrieval. It also modernizes the podcast generation pipeline by replacing legacy provider strings with a model registry system, allowing episode and speaker profiles to reference LLM and TTS models by record ID and properly passing configuration parameters like max\_tokens through to the generation services.
_open\notebook/podcasts · high confidence
Removal of legacy prompt templates
The application has removed four existing Jinja2 prompt templates: ask\_content.jinja, chat.jinja, spr.jinja, and summarize.jinja. These files previously defined the system instructions and context structures for content querying, conversational chat with tool usage, Sparse Priming Representation generation, and incremental summarization. Their deletion indicates a shift in how the model interactions are configured, likely moving towards a new prompt management strategy or different underlying logic.
prompts · high confidence
Search page handles URL parameters and opens matched records directly
The search page now supports navigation via URL parameters (e.g., from the command palette), automatically triggering a search or ask based on the 'q' and 'mode' query strings. Additionally, when a user clicks on a search result, the system now opens the specific matched record (such as an insight) rather than its parent source, providing more precise access to the relevant content.
frontend/src/app/(dashboard)/search · high confidence
Settings form now gates Docling features on backend availability
The SettingsForm component now conditionally enables or disables Docling-related controls (OCR, formulas, vision) based on the backend's capability probe. When Docling is unavailable, the OCR toggle is disabled and hint messages are shown; when available, the controls are enabled. The form also includes tests to verify this gating behavior, including handling of loading states and error conditions.
frontend/src/app/(dashboard)/settings/components · high confidence
Standardized math formatting in transformation prompts
The transformation prompt template now explicitly instructs models to use $$...$$ for display math and $...$ for inline math, ensuring formulas render correctly in the UI. This change also clarifies that fenced LaTeX code blocks should only be used when the user explicitly requests the source code, preventing unnecessary code blocks in standard mathematical explanations.
prompts/transformation · high confidence
Transformations workspace: new UI, per-transformation model selection, and cleared-description fix
The Transformations page has been rebuilt with a new collapsible card-based list, a dedicated playground for testing transformations, and a default prompt editor. Users can now select a specific model for each transformation in the editor and playground, enabling per-transformation model overrides. The editor also fixes a bug where clearing the description was ignored; it now persists empty descriptions correctly. The playground renders output with Markdown, tables, and LaTeX math.
frontend/src/app/(dashboard)/transformations · high confidence
Fixes
Enable end-to-end SSE streaming via Next.js API proxy
Added a new internal SSE proxy utility (\_sse-proxy.ts) and two Next.js API route handlers (search/ask and chat messages) that forward requests to the upstream backend while preserving the Server-Sent Events stream. This ensures that real-time streaming responses are correctly passed through the Next.js frontend layer without buffering or interruption, improving the responsiveness of search and chat interactions.
frontend/src/app/api · high confidence
Test coverage
Added test coverage for core application behaviors; Added test environment setup and mocks.
Dependencies
Migrate Python build system from Poetry to uv and setuptools
The project has switched its Python dependency management from Poetry to uv, replacing the \poetry.lock\ file with a \pyproject.toml\ configured for setuptools and uv. This migration introduces dependency groups for development and notebook environments, updates the Python version constraint to \\>=3.11,\<3.13\, and adds security constraint dependencies (such as \anyio\, \pyjwt\, and \urllib3\) to prevent the installation of versions with known vulnerabilities.
(dependencies) · high confidence
Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.
How this codebase got here
This is the PUBLIC form of this artifact. Findings are listed in full, but the details of SECURITY findings — which rule fired, in which file, on which line, and how to fix it — are deliberately withheld, and any secret-scanner results are excluded entirely. Where detail is absent here it was REMOVED FOR PUBLICATION; it is not missing from the analysis. The complete artifact is available from the repository owner.
Score
- CAI 53 → 47 (-5.9)
- Rubric changed (rubric-2026.10.1 → rubric-2026.10.5) — scores are not directly comparable.
Lenses
- Code Health 55 → 54 (-0.2)
- Architecture 61 → 61 (+0.0)
- Maturity 72 → 47 (-24.8)
- Readiness 56 → 51 (-4.5)
- Security 79 → 73 (-6.0)
- Accessibility 46 → 41 (-4.9)
- Performance 100 → 100 (+0.0)
Resolved (4)
- ADR not followed: ADR-011: Visual identity is a token contract in globals.css, reviewed through /dev/design (docs/7-DEVELOPMENT/decisions/ADR-011-design-token-system.md)
- Documentation: no installation or build instructions (docs/0-START-HERE/index.md)
- Medium CVE: [GHSA redacted] (frontend/package-lock.json)
- PR-triggered workflow grants a write-scoped token
New (15)
- ADR not followed: ADR-013: ObjectModel.get raises NotFoundError only for a missing record (docs/7-DEVELOPMENT/decisions/ADR-013-objectmodel-get-error-contract.md)
- AddSourceDialog.AddSourceDialog::isStepValid (cognitive 22) (frontend/src/components/sources/AddSourceDialog.tsx)
- AddSourceDialog.AddSourceDialog::isStepValid (cyclomatic 17) (frontend/src/components/sources/AddSourceDialog.tsx)
- Critical CVE: [GHSA redacted] (uv.lock)
- Critical vulnerability: [GHSA redacted] (uv.lock)
- High CVE: [GHSA redacted] (frontend/package-lock.json)
- High CVE: [GHSA redacted] (frontend/package-lock.json)
- High CVE: [GHSA redacted] (frontend/package-lock.json)
- High vulnerability: [GHSA redacted] (uv.lock)
- Leaked secret: hardcoded-credential (.github/ISSUE_TEMPLATE/installation_issue.yml)
- Medium CVE: [GHSA redacted] (frontend/package-lock.json)
- Medium CVE: [GHSA redacted] (uv.lock)
- Medium: security finding (details withheld)
- Orphaned file (frontend/src/components/common/LanguageLoadingOverlay.tsx)
- PR-triggered workflow grants a write-scoped token (.github/workflows/build-dev.yml)
Architecture
- 0 containers · 1 bounded contexts · 0 dependency edges (baseline)
Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.
Survey your own repository
lfnovo/open-notebook was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.
About this page
- The score is its most recent published measurement, taken on 21 October 2026 at a pinned commit. It is not a live figure and does not change until the project is measured again.
- Measured at commit a0f6f3081ae06328f0a451b3a3ef2959ca25721d — the exact code this score is about.
- Scored under rubric-2026.10.5 — the same rubric and the same method as every other entry in this index.
- Measured by watchdog.canine.dev using codehealth-analyzer measure-2026.10.3+preprod-fddbc92cb8df.