AxonHub is a production-grade, multi-provider LLM API gateway that routes, transforms, and manages requests to a wide range of AI providers (OpenAI, Anthropic, Gemini, DeepSeek, xAI, and 60+ others). It provides a unified API surface for chat completions, embeddings, moderations, audio, images, and video generation, with sophisticated features like load balancing (latency-aware, weighted round-robin, circuit breaker), per-channel rate limiting, provider quota enforcement, streaming, and reasoning token handling. The backend is written in Go using Gin, Ent ORM, gqlgen, and Uber Fx for dependency injection, supporting PostgreSQL, MySQL, and SQLite with automatic migration. The frontend is a comprehensive React/TypeScript SPA (Vite, TanStack Router/Query, Tailwind, shadcn/ui) that provides administration of channels, models, API keys, users, roles, projects, and system settings. It includes analytics dashboards, request tracing, usage logs, and OIDC SSO. The system is designed for multi-tenancy with fine-grained RBAC (scope-based permissions) and privacy policies enforced at the Ent ORM layer. Deployment is supported via Docker, Docker Compose, and Helm charts. The codebase exhibits strong engineering practices: extensive unit and integration tests (including race-condition tests and E2E Playwright tests), pipeline middleware for request transformation, generic caching abstractions, stream processing utilities, and a Zanzibar-inspired authorization framework. While there are minor issues (duplicate utility packages, a typo in a folder name, and some frontend boilerplate repetition), the overall architecture is clean, modular, and well-documented. The project is actively maintained with frequent commits, issue triage, and a clear release process.
LiteLLM is a production-grade, open-source proxy/gateway for over 200 large language model (LLM) providers, including OpenAI, Anthropic, Azure, Bedrock, Gemini, and many more. It presents a unified OpenAI-compatible API, handling request routing, authentication, rate limiting, cost tracking, caching, guardrails (content filtering, PII detection), and audit logging. The system is architected as a FastAPI Python server with a Next.js admin dashboard, Rust extensions for performance-critical paths (token counting, caching, auth, cost calculation), and optional PostgreSQL/Redis backends. It supports streaming, MCP (Model Context Protocol), A2A (Agent-to-Agent) protocol, and enterprise features like SSO, secret management, and Terraform deployment modules. Key entry points include the proxy server (litellm/proxy/proxy_server.py), CLI commands (litellm, litellm-proxy), and the admin UI at /ui/litellm-dashboard. The primary data flow is from client requests (chat completions, embeddings, audio, images) received by the proxy, which authenticates, applies guardrails, routes to the configured LLM provider(s), logs usage, and returns a standardized response. Caching is multi-tier (Redis, disk, S3, GCS, Azure Blob, semantic). Spend tracking is aggregated into daily tables and exposed via the management API and UI. External dependencies include PostgreSQL (via Prisma ORM), Redis (for distributed caching, rate limiting, and queuing), and cloud-specific secret managers (AWS Secrets Manager, GCP Secret Manager, Azure Key Vault, HashiCorp Vault). The project uses uv for Python dependency management, Cargo for Rust crates, and npm/pnpm for the dashboard. Extensive CI/CD pipelines (CircleCI, GitHub Actions) run unit, integration, e2e, mutation, and Rust-Python parity tests.
OGX (formerly Llama Stack) is a production-grade, open-source AI inference platform that provides a drop-in replacement for the OpenAI API. It is a multi-provider API gateway supporting chat completions, embeddings, the Responses API (agentic turns with tool use), vector stores, file processing, conversations, prompts, skills, and connectors. The system is built on a clean provider plugin architecture where inline (built-in) and remote providers (30+ including Anthropic, Bedrock, Gemini, Ollama, vLLM, NVIDIA) are registered via ProviderSpec and dynamically routed. The FastAPI server auto-discovers API protocols and generates routes with telemetry (OpenTelemetry), authentication, multi-tenancy, and ABAC access control. The repository includes a Next.js admin UI, a Python CLI, OpenAPI-generated client SDKs (Python, TypeScript), benchmarking tools (RAG eval, vertical scaling), and comprehensive integration tests with recording/replay. It is designed for self-hosted or cloud deployment with Kubernetes and Docker support.
genblaze is a production-grade Python monorepo SDK for orchestrating generative AI media pipelines — image, video, audio, text-to-speech, and chat — across 16+ provider integrations (OpenAI, Google Gemini/Imagen/Veo, NVIDIA NIM, GMICloud, Runway, Luma, Replicate, FAL, ElevenLabs, AssemblyAI, Hume, LMNT, Stability Audio, Decart, AtlasCloud). The public API is a chainable Pipeline/Step abstraction with batch, streaming, ingest, and agent-loop modes, plus a Click-based CLI (extract/verify/replay/index) for manifest inspection, a schema-first spec package generating TypeScript types from JSON Schemas, and ~30 runnable examples. Architecturally it is a monorepo: libs/core (genblaze-core) contains the framework (pipeline, providers, model registry, storage, media handlers, canonical hashing, signing, observability, webhooks), while each connector under libs/connectors/* is an independently versioned package registering itself via plugin entry points. Key entry points are Pipeline.run()/batch_run()/stream() for programmatic use and the genblaze command line; data flows through provider submit/poll/fetch_output lifecycles, optional caching, S3/B2 object-storage sinks with URL policies, and manifest generation that can be embedded directly into output media files. Two capabilities set this apart. First, provenance: every run produces a canonical, hash-verified Manifest (SHA-256 over deterministically serialized JSON), optionally Ed25519-signed, and byte-preservingly embedded into JPEG/PNG/WebP/MP4/audio files via XMP/atom splicing without re-encoding. Second, security and reliability engineering: SSRF protection with DNS pinning, credential redaction, ReDoS-safe pattern heuristics, ffmpeg injection escaping, configurable retry policies, model liveness probes, and discovery-aware validation (NATIVE/PARTIAL/NONE support levels) that reduces hardcoded slug rot. The project is actively maintained with unusually thorough tooling: golden hash-stability tests, a provider conformance test base class (~15 contract checks per connector), pin-parity and PyPI-metadata release gates, zizmor/dependabot CI, and live probe scripts that fail on API drift. Main risks are the inherent churn of 16+ third-party APIs (mitigated but not eliminated by mock-heavy tests) and a package/versioning layout complex enough that installs rely on documented make targets. Overall this is a strong, well-architected application SDK — suitable both for direct use and as a reference for multi-provider integration frameworks.
Claw Code is a production-grade CLI application for AI-assisted coding, built primarily in Rust with a complementary Python parity workspace. It serves as a local agent harness that integrates with multiple LLM providers (Anthropic, OpenAI, xAI, DashScope, local OpenAI-compatible), manages conversational sessions, enforces workspace-scoped permissions, supports the Model Context Protocol (MCP) for external tool integration, and provides a plugin system for lifecycle hooks. The architecture is a multi-crate Rust workspace under `rust/`, with the main binary `claw` (from `rusty-claude-cli`), a core runtime crate (`runtime`) handling session persistence, permission enforcement, conversation loops, and MCP plumbing; an API client crate (`api`) for provider dispatch and streaming; a plugins crate for lifecycle management; and supplementary crates for tools, telemetry, and a workspace RAG service. The Python workspace (`src/`) is explicitly a non-production porting/parity mirror of a TypeScript archive, used for surface mapping and audit. The Rust code is well-structured, extensively tested, and demonstrates strong engineering practices—clean abstractions, reusable components, and defense-in-depth security (path validation, read-only command heuristics, approval tokens). Notable highlights include a permission enforcer with workspace boundary enforcement, a modular MCP client supporting multiple transports, a plugin system with manifest validation, and a session store with JSONL rotation and workspace fingerprint isolation. The repository is actively maintained (CI workflows, issue triage, recent commits) and shows low technical drift, though some modules have `#[allow(dead_code)]` annotations and the parity workspace adds maintenance overhead.
gpt4free (g4f) is a large-scale, community-driven Python library that aggregates dozens of free AI providers (LLMs, image, audio, video) into a unified, OpenAI-compatible interface. It provides multiple access modes: a Python client library (sync/async), a FastAPI-based API server, a Flask web GUI, a CLI tool, a Model Context Protocol (MCP) server for tool calling, a browser extension, a Discord bot, an Android app, and a Go launcher. The architecture is provider-based: each AI service (e.g., ChatGPT, Gemini, DeepSeek, Copilot, Grok, Perplexity) is wrapped in an adapter class implementing a common `AsyncGeneratorProvider` interface. The system handles authentication via cookies, OAuth flows, HAR files, and browser automation (CDP/nodriver), and supports streaming, tool/function calling, image generation, audio synthesis, and web search. It is primarily an application for consuming free AI models, not a research or data pipeline project, though it includes some data-oriented utilities like token optimization and usage tracking.
One API is a production-ready OpenAI API gateway that aggregates 30+ LLM providers (OpenAI, Anthropic, Google Gemini, Azure, Baidu, Alibaba, Zhipu, Moonshot, DeepSeek, Ollama, etc.) behind a single OpenAI-compatible endpoint. The Go backend implements a sophisticated relay system with provider-specific adaptors that translate OpenAI-format requests to each provider's native API. It features token-based quota management with group/channel multipliers, load balancing across channels, automatic failover, channel health monitoring with metric-based auto-disable, and multi-node deployment with Redis synchronization. The React frontend (three themes: default, berry, air) provides admin dashboards for channel/token/user management, billing visualization, and system configuration. The codebase demonstrates strong engineering practices: clean adaptor pattern for provider extensibility, comprehensive middleware chain, batched database updates for performance, and well-structured configuration management.
This is a production-grade, AI-powered code review automation tool that integrates with multiple version control systems (GitHub, GitLab, Bitbucket Cloud/Server, Azure DevOps, Gitea) and multiple LLM providers (OpenAI, Claude, Gemini, Ollama, Bedrock, OpenRouter, Azure OpenAI). The system is implemented as a Python async CLI application, packaged as a Docker image, and includes CI/CD integration templates for GitHub Actions, GitLab CI, Jenkins, Bitbucket Pipelines, and Azure Pipelines. Architecturally, it follows a clean layered design: a CLI layer (Typer) drives commands; a configuration layer (Pydantic Settings with multiple file/env sources) manages nested settings for LLM, VCS, agent, review, artifacts, and logging; a service layer orchestrates the review process through runners, gateways, and internal services (prompt building, diff analysis, policy enforcement, cost tracking, artifact persistence); and a client layer provides unified abstractions for VCS and LLM providers with provider-specific implementations. The tool supports multiple review modes: inline comments on specific code lines, summary comments on the entire PR, context review (full repository exploration via an optional agent loop), and reply to existing comments. The agent loop uses a ReAct pattern with tool execution (shell commands like `rg`, `cat`, `git`) to gather deeper context, with duplicate detection and context limits. A dry-run mode allows testing without posting to VCS. The system also includes a hook service for lifecycle events, cost tracking per request, and optional artifact saving of LLM interactions. Key strengths include extensive test coverage (unit tests with fake clients and fixtures), generic reusable components (HTTP client with retry, pagination helpers, diff parser, LLM output JSON parser, cost service), and a well-structured configuration system that supports discriminated unions for provider-specific settings. The codebase is actively maintained, uses modern Python async patterns, and demonstrates high craftsmanship in its abstraction layers and separation of concerns.
NextChat is a production-grade, multi-provider AI chat client built as a Next.js web application with an optional Tauri desktop wrapper. It supports 15+ LLM providers (OpenAI, Anthropic, Google, Baidu, DeepSeek, etc.) through a unified API abstraction layer, and provides a rich chat interface with streaming responses, markdown rendering, code folding, image upload, TTS, real-time voice chat, and Stable Diffusion image generation. The application is designed for self-hosting (Docker, Vercel, manual) and emphasizes privacy by storing all chat data locally in IndexedDB/localStorage, with optional cloud sync via WebDAV or UpStash Redis. Architecturally, NextChat is a monolith with client-side state management using Zustand (with IndexedDB persistence), serverless API routes for proxying LLM requests, and a React component library. It includes a plugin system for OpenAPI-based function calling, MCP (Model Context Protocol) integration for external tool execution, and a prompt/mask system for reusable system instructions. The codebase is well-structured with reusable patterns like generic store factories, UI components, and streaming utilities, though it exhibits significant duplication across provider integrations and some features (e.g., Tencent, SD) appear less maintained.
Vagrant is a mature, widely-used CLI tool for building and managing portable development environments. It orchestrates virtual machines and containers across multiple providers (VirtualBox, Hyper-V, Docker) with a plugin-based architecture. The system reads Vagrantfile configurations, executes a middleware action chain to manage machine lifecycle (up, provision, halt, destroy, snapshot, package), and supports guest OS-specific capabilities (network configuration, shared folders, hostname changes) via a capability host pattern. It includes built-in provisioners (Shell, Chef, Puppet, Ansible, Salt, Docker, Podman, File), communicators (SSH, WinRM, WinSSH), synced folder backends (NFS, rsync, SMB), and cloud integration with Vagrant Cloud. The codebase is predominantly Ruby with a newer Go-based plugin SDK. It has extensive unit and acceptance tests, a well-defined plugin API (V1/V2), and reusable utility modules. While the core is solid, the codebase shows signs of technical drift with deprecated features, version-specific workarounds, and a dual-language complexity.
Chatbox is a cross-platform desktop AI chat client (Windows/macOS/Linux) built with Electron, React, and TypeScript. It supports 30+ AI providers (OpenAI, Claude, Gemini, DeepSeek, Ollama, etc.) via a unified provider registry and model abstraction layer. The application features agent mode with tool use, sandboxed code execution, MCP server integration, knowledge base RAG, web search, image generation, and a skills marketplace. It uses a hexagonal architecture with ports/adapters in a core package (`@chatbox/core`) that is host-agnostic, and React bindings in `@chatbox/react`. The renderer uses TanStack Router, Zustand/Jotai for state, and Mantine/MUI for UI. The main process handles IPC, OAuth flows, file system sandbox, and analytics. The codebase is large and well-tested (Vitest unit tests, integration tests, contract tests) with a focus on offline-first, local storage, and cross-platform compatibility (desktop, web, mobile via Capacitor). The application is actively maintained with a commercial 'Pro' version that shares code with this open-source Community Edition. Notable engineering patterns include a clean port/adapter separation, in-memory test doubles for all repositories, a session write coordinator that serializes updates and cancels stale queries, a draft tokenization system with web workers, and a provider registry that allows dynamic model discovery. The codebase has some technical debt: multiple UI libraries, legacy migration paths, and commented-out code, but overall demonstrates strong craftsmanship and reusability of its core abstractions.
Graphiti is a Python library and server framework for building temporal knowledge graphs that serve as persistent memory for AI agents. It ingests unstructured text episodes (messages, transcripts, JSON), uses LLMs to extract entities and relationships with bi-temporal validity windows, deduplicates against existing graph data, and provides hybrid search (semantic + BM25 + graph traversal) with configurable reranking. The codebase is structured around a clean driver abstraction (Neo4j, FalkorDB, Kuzu, Neptune) with per-provider operations classes, a prompt library supporting versioning and overrides, and a multi-provider LLM/embedder/cross-encoder client system. It is primarily an application library, though it includes evaluation infrastructure and research-grade deduplication algorithms (MinHash/LSH).
Raif (Ruby AI Framework) is a Rails engine/gem that provides a comprehensive, opinionated framework for integrating large language models (LLMs) into Rails applications. It abstracts over multiple LLM providers (OpenAI, Anthropic, Google Gemini, AWS Bedrock, xAI, OpenRouter) with a unified interface for chat completions, streaming, tool calling, structured outputs, and batch inference. The system persists every LLM interaction as `ModelCompletion` records with full token/cost accounting, supports conversation management with tool call validation and retry loops, and provides a task system for running parameterized LLM prompts with caching and temperature control. A key differentiator is its built-in evaluation framework (`Raif::Evals`) that enables dataset-driven evals with LLM-as-a-judge, statistical comparison (bootstrap CI95, sign test, Fisher exact), and regression gating for CI/CD pipelines. Architecturally, Raif is a mountable Rails engine with its own models, migrations, controllers, background jobs, and admin UI. It uses Single Table Inheritance extensively for extensibility (tasks, agents, model completion batches, LLM judges). The system includes a declarative model manifest DSL for defining model capabilities, pricing, and lifecycle, a data retention/archiving subsystem with pluggable storage adapters, and a Prompt Studio for A/B testing prompts across models. The eval framework supports concurrent execution via forked worker processes with per-worker databases, live HTML reports, and resumable run logs. External dependencies include Faraday for HTTP, ActiveJob for background processing, and provider-specific SDKs (AWS SDK for Bedrock). Key entry points include `Raif::Llm#chat` for direct LLM calls, `Raif::Task#run` for task execution, `Raif::Conversation` for multi-turn conversations, `Raif::Agents::NativeToolCallingAgent` for ReAct-style agent loops, and the CLI (`raif evals`, `raif evals:compare`, `raif evals:report`) for running evaluations. The admin web interface provides full CRUD for tasks, conversations, agents, model completions, and batch inference jobs. Background jobs handle archive/cull operations, batch polling, and cost event repair. The system is designed for production use with features like advisory locks for concurrency control, transient retry with exponential backoff, and configurable data retention policies. While primarily an application framework, Raif has strong data/research characteristics through its evaluation framework, model manifest with pricing/capability tracking, and smoke-testing infrastructure for live API verification. The eval framework produces detailed metrics (pass rates, scores, token usage, costs, confidence intervals) and supports comparison reports with regression gates, making it suitable for systematic LLM evaluation in CI/CD workflows. The model manifest serves as a machine-readable catalog of supported models with lifecycle management, enabling automated capability detection and cost estimation.
MoneyPrinterTurbo is a production short-video generation platform. Given a topic or script, it orchestrates LLM-based script/keyword generation, multi-provider TTS (Edge, Azure, Gemini, ElevenLabs, etc.), video material sourcing (Pexels, Pixabay, Coverr, several AI video generators), subtitle generation (Whisper or TTS timestamps), background music, and final video composition via moviepy/FFmpeg. It exposes three consistent interfaces: a FastAPI REST API, a Streamlit WebUI, and a CLI. All three share the same underlying task pipeline in `app/services/task.py`, which manages state, progress, and cross-posting to YouTube/TikTok/Instagram. The architecture is a service-oriented monolith: the core pipeline is a linear sequence of stages (script → terms → audio → subtitles → materials → video → optional BGM/cross-post). External integrations are abstracted behind provider-specific modules (e.g., `voice.py`, `material.py`, `bgm.py`). A Redis-backed or in-memory state manager tracks task progress. The configuration is runtime-editable and stored in a TOML file. The WebUI is a large Streamlit single-page application with dozens of provider-specific controls, billing confirmations, and i18n support for 12+ languages. The codebase demonstrates strong engineering practices: extensive test coverage (200+ test files), structured error handling with per-provider error classes, atomic file writes, path traversal protection, and security-conscious API key handling. However, it also shows signs of rapid growth: the WebUI is a monolithic 1300+ line script, provider integration logic is occasionally duplicated, and the dependency list is large (many third-party APIs). The repository is actively maintained with recent commits and good issue triage. It is not a research project; there are no datasets, model training, or analytical queries.
AnythingLLM is a full-stack, self-hosted AI chat platform designed for Retrieval-Augmented Generation (RAG) over private documents. It provides a web interface where users create workspaces, upload documents (PDF, DOCX, images with OCR, audio, etc.), which are then chunked, embedded, and stored in a vector database (default LanceDB, with support for PGVector, Pinecone, Qdrant, and others). Chat queries retrieve relevant document chunks and feed them to an LLM of choice (supporting 30+ providers including OpenAI, Anthropic, Ollama, and local models via KoboldCPP). The platform also features an extensible agent system (aibitat) with plugins for file system operations, SQL databases, Gmail, Google Calendar, web scraping, image generation, and scheduled jobs. A separate 'collector' microservice handles document ingestion and format conversion, while the 'open-computer' subproject provides a QEMU-based virtual machine for autonomous browser automation. The frontend is a React SPA with multi-user support, role-based access, internationalization, embeddable chat widgets, and a community hub for sharing agent skills, system prompts, and slash commands. The architecture is a monolith (Express server) with the collector as an isolated service, and the open-computer as a standalone CLI/service. It relies on external AI APIs, vector databases, and local storage, making it suitable for both cloud and on-premises deployment.
Tachikoma is a production-grade Swift SDK/library that provides a unified typed interface to a broad spectrum of LLM providers: OpenAI (Chat Completions, the newer Responses API for GPT-5+ generations, and a ChatGPT Codex OAuth backend), Anthropic (Claude with streaming signed thinking blocks), Google Gemini, Mistral, Groq, Grok/xAI, Kimi, MiniMax, OpenRouter, Together, Replicate, Ollama, LM Studio, Azure OpenAI, and arbitrary OpenAI/Anthropic-compatible endpoints. Its public surface is deliberately modeled on the Vercel AI SDK: generateText/streamText, generateObject/streamObject with partial-JSON repair, embeddings, vision analysis, and multi-step agent tool-calling loops. Beyond generation, it ships composable stop conditions, response caching, retry policies, usage/cost accounting, reasoning-effort controls, and a provider feature-parity adapter that normalizes differences in system roles, vision, and tool limits across backends. The repo is strictly an application SDK in character: no datasets, notebooks, or data pipelines. The one data-like facet is its manually curated model capability catalog (context lengths, pricing, per-generation reasoning support), which is embedded as Swift enums rather than external registries. Architecturally it is a SwiftPM monorepo of five products: Tachikoma (core generation/provider layer), TachikomaAgent (conversations, dynamic tool registry, MCP tool binding), TachikomaAudio (transcription, TTS, and an OpenAI Realtime WebSocket voice client), TachikomaMCP (a client for Model Context Protocol servers over stdio/SSE/HTTP with full JSON Schema bridging), and TachikomaConfigCLI (auth/credential helper), plus two example CLIs (ai-cli, gpt5cli). All providers implement a common ModelProvider protocol with ProviderRequest/ProviderResponse and AsyncThrowingStream-based streaming, with shared SSE/NDJSON parsing funneled through OpenAICompatibleHelper and provider-specific adapters for Anthropic, Google, and Ollama. The code is Swift 6.2, heavily async/await, and covered by a substantial test suite (Swift Testing, mock URLProtocol fixtures, snapshot tests, serialized mutex-guarded tests for shared state), with CI across macOS and Linux. The codebase exhibits strong engineering practice: clean provider abstraction, genuinely reusable utilities (stream transforms, retry policies, TTL caches, typed JSON Schema views), modular product boundaries, and disciplined error modeling. Its principal weaknesses are the enormous provider surface that must be manually maintained (hardcoded capability matrices, context-length string heuristics like isOpus48/isSonnet5, static pricing tables in ModelCostCalculator), a handful of @unchecked Sendable singletons and NSLock-guarded global state, Linux/Apple streaming divergence (Linux buffers the full response instead of streaming bytes), some dead placeholder APIs (the legacy generate() returning a mock string), and a rapid feature-churn cadence visible in the changelog (GPT-5.6/6, Claude 5, reasoning replay). These are maintainability risks more than correctness defects; the tested behavior is consistently thorough. Overall this is a high-craft SDK that an engineer building AI application features in Swift would benefit from studying, particularly its OpenAI-compatible adapter pattern, cache-key determinism, and MCP schema translation.
This repository is an open-source terminal-based AI coding assistant built in Rust, clean-room reimplementing Anthropic's Claude Code. It provides an interactive TUI (ratatui) for multi-provider LLM conversation (Anthropic, OpenAI, Google, and 30+ OpenAI-compatible) with a rich tool ecosystem for file editing, bash execution, MCP integration, code analysis, web fetching, and agent orchestration. The architecture is modular: a provider abstraction layer (`LlmProvider` trait), plugin system, slash commands, session management, and an Agent Client Protocol (ACP) for external IDE integration. The codebase is organized into 12+ workspace crates with clear separation of concerns, but many modules are still in active refactoring (issue #232) with stubs, TODOs, and incomplete feature gates (36 experimental features). The permission system, tool budget enforcement, and streaming decoders demonstrate strong engineering, while dual feature flag systems and global mutable state introduce maintenance risk. Overall, it is a feature-rich, well-architected application with some rough edges from ongoing migration.
KEA Research is a full-stack multi-provider AI research chat application. The backend is a FastAPI service backed by SQLite/SQLAlchemy, exposing chat, admin, health, user, and TTS endpoints. The frontend is an Astro + TypeScript single-page app that stores chat history and pipeline data in IndexedDB, with a Bootstrap-based UI. Deployment is containerized via Docker Compose with nginx handling TLS and SSE-friendly proxying. The core product logic is a four-step research pipeline: initial multi-provider responses, Mixture-of-Agents refinement, peer evaluation, and a final KEA synthesis with facts, consensus, flagged claims, rankings, and confidence scores. A web admin panel manages providers, models, provider sets, users, and app settings, while the frontend supports research layers, bookmarks, notes, drafts, export, attachment ingestion, and text-to-speech via a Kokoro ONNX model.