The OpenAI Agents SDK (Python) is a production-grade library for building agentic AI applications. It provides a complete agent runtime (`Runner`, `Agent`, `RunConfig`) with support for tool calling, handoffs, guardrails, sandboxed execution (Docker, Modal, E2B, Cloudflare, Vercel, Unix-local), realtime voice/audio pipelines, MCP integration, and multi-provider model routing (OpenAI, LiteLLM, AnyLLM, custom). The codebase is async-first with extensive tracing/observability, session persistence (SQLite, OpenAI Conversations, Responses Compaction), and a rich lifecycle-hook system. It includes a large test suite covering unit, integration, and sandbox security scenarios, plus internal tooling skills for release management and code review. The SDK is actively maintained with a 0.23.x version, release automation via release-please, and comprehensive docs in English and Japanese.
FastMCP is a production-oriented Python framework for the Model Context Protocol (MCP). Rather than being a single application, it is a multi-package library stack: `fastmcp_slim` provides the core server/client framework, `fastmcp_remote` bridges remote MCP servers to local stdio, and `fastmcp_tasks` implements SEP-2663 background task execution. The core architecture centers on the `FastMCP` server class, a `Client` class with transport abstraction (stdio, SSE, Streamable HTTP, in-memory), provider-based component registration (local, filesystem, OpenAPI, proxy, skills), composable middleware, transforms, OAuth/OIDC authentication, OpenTelemetry instrumentation, and a CLI for running, installing, deploying, and inspecting servers. Entry points include the `fastmcp` CLI and the Python APIs `FastMCP`, `Client`, `ClientGroup`, `FastMCPApp`, and `TasksExtension`. This is an application framework, not a data/research repository, though it contains information-retrieval components (BM25 and regex tool search), OpenAPI schema conversion machinery, and telemetry utilities that approach data-engineering utility. The codebase demonstrates strong engineering discipline: extensive unit and integration tests, conformance testing against the MCP spec, lazy imports to preserve startup performance, upgrade-guide tests that verify documentation claims, and a transform/middleware system that keeps extensions decoupled from core server logic. External dependencies include the MCP Python SDK, Pydantic, Starlette/ASGI, OpenTelemetry, optional authlib, optional prefab-ui for interactive apps, and an internal `httpx2` fork, which is the most notable maintenance risk in an otherwise current codebase.
MLflow is an open-source ML lifecycle platform that provides experiment tracking, model registry, model deployment, evaluation, and observability for both classical ML and GenAI/LLM workflows. In production, it operates as a modular monolith: a Python backend (Flask/FastAPI) with SQLAlchemy-backed storage (SQLite, PostgreSQL, MySQL, MSSQL), a React/TypeScript web UI, and client SDKs for Python, R, Java, and TypeScript. The system exposes a REST API and CLI for logging experiments, registering models, collecting traces via OpenTelemetry, and managing an AI Gateway that routes requests to dozens of LLM providers with budget controls, guardrails, and traffic splitting. The architecture includes a plugin system for custom store backends, deployment targets, and authentication providers, as well as a custom linting tool (clint) and extensive CI/CD pipelines. Key entry points are the `mlflow` CLI, `MlflowClient` Python API, tracking server (`mlflow server`), gateway server (`mlflow gateway`), and model scoring server (`mlflow models serve`). Data flows from user instrumentation through asynchronous logging to the backend store, with traces and evaluations queryable via the UI or API. External dependencies span ML frameworks (scikit-learn, PyTorch, TensorFlow, LangChain, etc.), LLM provider SDKs, and infrastructure tools (Docker, Kubernetes, Helm).
Composio is a production-grade, dual-language SDK monorepo that connects AI agents to external services (Gmail, GitHub, Slack, Notion, Salesforce, and hundreds more) through a unified tool abstraction. In production it operates as the client layer of the Composio platform: applications create a Composio session bound to a user_id, discover toolkits, manage OAuth auth-configs and connected accounts, and execute tool calls either remotely or via a session-scoped Tool Router that also exposes hosted MCP endpoints. The repository ships a TypeScript SDK (@composio/core), a Python SDK (composio), a standalone Effect-TS-based CLI, 15+ framework provider adapters (OpenAI, Anthropic, LangChain, LlamaIndex, CrewAI, AutoGen, Mastra, Vercel AI, Google/ADK, etc.), and a Next.js documentation site. It is unambiguously an application/developer-platform repo rather than data research, although it contains data-adjacent engineering assets: a cross-SDK JSON Schema conversion corpus, OpenAPI dereferencing tooling, a knowledge-base hybrid-search index, and telemetry redaction utilities.
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.
AIStore (AIS) is a high-performance, distributed object storage system engineered for AI/ML workloads. It operates as a cluster of proxy (gateway) and target (storage) nodes, providing S3-compatible API, multi-cloud backend support (AWS, GCP, Azure, OCI), and advanced data management features such as erasure coding, mirroring, LRU caching, and distributed resharding (dsort). The system is designed for linear scale-out and consistent performance, with a rich set of extended actions (xactions) for batch operations like copy-bucket, archive, blob download, and ETL transformations. It includes a full-featured CLI (`ais`), a Python SDK with PyTorch dataset integration, and a standalone authentication server (AuthN). The codebase is primarily a production-grade application with strong software engineering practices: modular architecture, extensive unit and integration tests, code generation for high-performance serialization (msgp), and build-tag-based feature gating. It demonstrates non-obvious craftsmanship in areas like memory management (slab allocators, SGL), intra-cluster transport (stream bundles), and probabilistic data structures (dynamic cuckoo filters). While the core is a storage system, it has significant data/research features (ETL pipelines, distributed sorting, shard indexing) that make it a hybrid tool for AI data pipelines. Key entry points include `cmd/aisnode/main.go` (server), `cmd/cli/main.go` (CLI), `cmd/authn/main.go` (authentication), and `python/aistore/sdk/` (Python SDK). The system relies on cloud SDKs, Kubernetes client, and various Go libraries. It is actively maintained with frequent commits, comprehensive CI/CD, and a large test suite.
maxapi is an async Python SDK and bot framework for the MAX messenger platform. It provides a high-level client (`Bot`) for all MAX Bot API endpoints, an event-driven dispatcher with router hierarchy (`Dispatcher`/`Router`), composable filters (magic filter `F`, `Command`, `StateFilter`, etc.), middleware chains, a finite state machine (FSM) with multiple backends (memory, Redis), and webhook integrations for aiohttp, FastAPI, and Litestar. The library is heavily inspired by aiogram but adapted to MAX API specifics such as attachment types, callback payloads, and deep linking. It includes extensive utility code for formatting (HTML/Markdown), inline keyboard building, VCF parsing, and time helpers. The codebase is well-structured with Pydantic v2 models for all API entities, rigorous retry logic, and a comprehensive test suite covering unit, integration, and webhook scenarios. While not a standalone application, it serves as a foundational component for building production-ready MAX bots.
VieNeu-TTS is an advanced on-device Vietnamese text-to-speech library with instant voice cloning, supporting bilingual (Vietnamese/English) synthesis. The system provides a Python SDK (`vieneu`) with multiple backends: v3 Turbo (default, 48kHz, ONNX or PyTorch), v3 Nano (lightweight 24kHz, ONNX only), and legacy v1/v2 (deprecated). It includes a Gradio web UI (`vieneu-web`), a FastAPI streaming server, and a browser-based HTML/JS client. The core architecture is a factory pattern (`Vieneu(mode=...)`) that returns the appropriate engine, all inheriting from a shared base class. Inference can run entirely torch-free on CPU via ONNX Runtime, with optional GPU acceleration (CUDA) for batched generation and lower latency. Feature highlights include instant voice cloning from a 3–8 second reference clip using a speaker encoder (ONNX) and denoiser, inline emotion cues (`[cười]`, `[thở dài]`), frame-level streaming, SRT subtitle-to-speech, and LoRA fine-tuning for custom voices (requires ~6GB GPU). The system is deployed via Docker with CPU and GPU profiles, and has CI/CD pipelines for testing and PyPI publishing. The codebase is well-structured with clear separation of concerns, though it carries some legacy weight from older model versions and a few deprecated parameters. Other notable components include a sliding-window repetition penalty, a babble guard heuristic for short utterances, and a voice persistence system for user-defined presets.
PageIndex is a production-ready RAG framework that replaces vector databases with a reasoning-based, hierarchical tree index over documents. Its core innovation is the 'flash' pipeline: a sophisticated, LLM-free PDF structure extractor that uses character-level layout analysis, trie-based keyword matching, and heuristic classification to build a table of contents tree from PDFs in seconds. The system then enables agentic retrieval by feeding this tree to an LLM, which reasons over the hierarchy to find relevant pages without embedding or chunking. The repository provides a full Python SDK (PageIndexClient) with local and cloud modes, CLI tool, integrations with OpenAI Agents SDK, Anthropic SDK, and Claude Agent SDK, plus an MCP bridge for remote tool execution. While primarily an application, it contains a deep research-grade data processing pipeline (flash) with extensive multilingual dictionaries and layout analysis heuristics, giving it a hybrid character. The codebase is well-structured with clear module boundaries, extensive test coverage, and reusable patterns like the tool contract system and trie-based matching. However, it carries medium technical drift risk due to hardcoded thresholds, dependency on private APIs of LLM frameworks, and pdfium version sensitivity.
TqSdk is a Python SDK for quantitative trading in Chinese futures, options, and stock markets. It provides a unified async event-driven framework for real-time market data, historical data download, backtesting, simulation, and live trading through multiple broker gateways (CTP, SimNow, etc.). The core architecture centers on a TqApi class that manages WebSocket connections, diff-based data synchronization, and a wait_update() loop. It includes a comprehensive technical analysis library (tafunc/ta), target position management (TargetPosTask), risk management rules, and a web-based GUI. While primarily an application library for building trading strategies, it also contains data analysis components like technical indicators and backtest report generation, making it a hybrid tool with a strong application focus.
AlphaGenome is a Python client library and SDK for interacting with Google DeepMind's AlphaGenome cloud service, a deep learning model that predicts molecular properties (gene expression, splicing, chromatin accessibility, histone marks, contact maps) from DNA sequence. The library provides gRPC-based clients for sequence, interval, and variant prediction, along with precomputed variant scores via the Atlas service. It includes rich data containers (TrackData, JunctionData), modular visualization components (tracks, transcripts, sashimi plots), and interpretation tools like in-silico mutagenesis. The repository also contains a set of Colab notebooks demonstrating end-to-end workflows for variant scoring, haplotype analysis, splicing quantification, and batch processing. The architecture is a thin client layer over a remote API, with protobuf definitions for service contracts and tensor serialization. Core modules handle genomic data types (Interval, Variant, Transcript), gene annotation parsing, ontology mapping, and efficient tensor packing. The variant scoring framework is extensible via dataclass-based scorers (center mask, gene mask LFC, splicing, polyadenylation) and includes utilities to tidy AnnData results into pandas DataFrames. Visualization is component-based, allowing composition of tracks, contact maps, and annotations into publication-quality figures. While primarily an application library, the repo has a strong data research character due to its focus on genomic analysis and the included notebooks. The code is well-structured, with extensive unit tests and CI/CD pipelines. However, it is tightly coupled to the AlphaGenome API and protobuf definitions, limiting standalone reuse. The most reusable components are the data containers (TrackData, JunctionData), tensor packing utilities, and the variant scorer dataclass pattern, which could be extracted for other genomic scoring tasks.
Langflow is a full-stack, open-source platform for building and deploying AI-powered agents and workflows. It provides a visual flow builder (React frontend), a Python backend (FastAPI/SQLModel), a CLI executor (LFX), an MCP server for AI tool integration, and a Python SDK. The codebase is a monorepo with multiple packages: the main `langflow` distribution, `langflow-base` (core backend), `lfx` (lightweight executor/CLI), `langflow-sdk` (API client), `langflow-stepflow` (step-based execution), and extension bundles (e.g., lfx-duckduckgo, lfx-arxiv, lfx-ibm, lfx-docling). The backend uses SQLModel with Alembic migrations, Celery for background tasks, FastAPI for REST endpoints, and a plugin/extension system for bundles. The frontend is a React/TypeScript app built with Vite, React Flow, Zustand, and Tailwind. The architecture supports multi-worker deployments, auto-login, JWT auth, rate limiting, MCP server integration, and a rich component library for LLMs, vector stores, tools, and data processing.
Mobilerun is a comprehensive CLI and Python SDK framework for controlling Android and iOS devices through natural language commands powered by LLM agents. It orchestrates multi-agent workflows (Manager, Executor, FastAgent) built on llama-index Workflows, with device interaction via ADB (Android) and a separate iOS binary (WebDriverAgent). The system supports multiple LLM providers (OpenAI, Anthropic, Gemini, xAI, DeepSeek, Ollama, etc.), including custom OAuth flows for subscription-based access (Anthropic Claude, Google Code Assist, xAI Grok). It features a configurable agent architecture with Jinja2 prompt templates, macro recording/replay, MCP-based custom tool integration, telemetry (PostHog), tracing (Phoenix, Langfuse), and a rich set of developer tools (device selection, configuration wizard, doctor, credential management). The codebase is well-tested with over 50 test files covering XML parsing, coordinate coordinate contracts, OAuth flows, and agent workflows. While there is some code duplication in OAuth implementations and tight coupling to llama-index internals, the overall architecture is modular and extensible. The project is in active development (version 0.7.0, Beta) with commercial ties to Mobilerun Cloud.
The Parcle Python client is a library for interacting with the Parcle Memory API, which provides long-term memory for AI agents. It enables developers to store conversational data and files per user, and then retrieve relevant information through natural language queries that return synthesized answers with citations. The client handles all aspects of communication with the API, including authentication, request retries, file uploads, and processing of Server-Sent Events for search operations. The architecture is straightforward: a synchronous HTTP client built on top of `httpx`, with methods corresponding to each API endpoint. Key features include robust error handling via a well-defined exception hierarchy, automatic retry logic for transient failures, and flexible file upload support that accepts various input types while enforcing size limits. The search method uniquely handles SSE streams to deliver incremental responses, ensuring compatibility with the API's streaming behavior. While the library is purpose-built for the Parcle API, it exhibits several reusable patterns. The exception mapping system in `exceptions.py` can be adapted for any HTTP API, and the file upload handling functions in `client.py` demonstrate a robust approach to processing diverse file inputs. The codebase is actively maintained, has low technical debt, and includes comprehensive tests that mock HTTP interactions to ensure reliability.
OpenBB is an open-source financial data platform that provides a unified, multi-interface access layer to over 20 data providers (FMP, FRED, CBOE, Yahoo Finance, Intrinio, etc.) through a modular provider-extension architecture. The core library (`openbb_core`) defines a standardized Fetcher pattern (transform_query -> extract_data -> transform_data), Pydantic-based standard models for equities, derivatives, fixed income, economics, and more, and a command routing system exposed via a Python SDK (OBBject), a REST API (FastAPI), a CLI (argparse translator), a desktop app (Tauri + React), and an MCP (Model Context Protocol) server for AI agents. The platform includes domain-specific extensions (equity, crypto, economy, quantitative, technical analysis, charting) and a cookiecutter template for scaffolding new providers or OBBject extensions. While the codebase is large and actively maintained, there is notable technical drift in the form of hardcoded API URLs, mixed sync/async extraction, deprecated endpoints, and code duplication across many provider implementations. The charting extension's plugin-based technical indicator system, the country/exchange validation utilities, and the LRU cache with TTL are high-craftsmanship, reusable components that demonstrate non-obvious software engineering skill.
POML (Prompt Orchestration Markup Language) is a developer toolchain for structured, reusable LLM prompt engineering. It provides an HTML-like markup language with semantic components, CSS-like styling, templating (variables, loops, conditionals), and multimodal data integration (PDFs, images, tables). The repository contains a core TypeScript/React SDK, a Python wrapper with OpenAI/LangChain/MLflow integrations, a VS Code extension (LSP, live preview, test command), and a Chrome browser extension for content extraction. It is packaged as CLI tools and libraries on npm and PyPI. Designed for building maintainable prompts in LLM applications, it also includes research artifacts (arXiv papers) and a tracing system for debugging prompt execution.
This repository implements the Agent-to-Agent (A2A) protocol in Python. It is a comprehensive SDK providing both server and client libraries for building interoperable agent systems. The architecture is modular with abstract interfaces (AgentExecutor, TaskStore, ClientTransport) and supports multiple transports: JSON‑RPC, REST/HTTP+JSON, and gRPC, with a full backward‑compatibility layer for v0.3. Key features include agent card signing/verification (JCS canonicalization), push notifications, event streaming, and database persistence via SQLAlchemy. The codebase is well-structured with extensive automated testing (unit, integration, cross‑version, and conformance), CI/CD pipelines, and release automation. While primarily an SDK for protocol compliance, it also includes a sample agent application and an interop test kit.
python-o365 is a mature Python library that provides a comprehensive, object-oriented wrapper around the Microsoft Graph API. It enables developers to programmatically access and manage Microsoft 365 services including Outlook email and calendar, OneDrive and SharePoint files, Teams chats and presence, Planner tasks, directory users and groups, and more. The library abstracts OAuth2 authentication (supporting authorization code, public client, and client credentials flows) and handles token refresh, pagination, and API versioning internally. It is designed as a client-side SDK, not a server application, and can be embedded into scripts, web applications (e.g., Flask), or automation workflows.
Stagehand is a Python SDK for the Stagehand API, an AI-driven browser automation framework. The SDK provides sync and async clients (`Stagehand`, `AsyncStagehand`) that interact with remote or local browser sessions via natural language commands like `observe`, `act`, `extract`, and `execute`. It supports both cloud-based (via REST API) and local execution modes, the latter using a compiled SEA binary managed by `SeaServerManager`. The codebase is primarily auto-generated by Stainless from an OpenAPI spec, with hand-written custom modules for session binding, binary management, and local server orchestration. The architecture follows a resource-based client pattern: the main client holds resource objects (e.g., `sessions`) that expose typed methods returning Pydantic models. Request parameters are defined as `TypedDict` subclasses with camelCase aliases, and responses are deserialized into strongly-typed models. The SDK includes comprehensive error handling (hierarchy of `APIError` subclasses), configurable retries and timeouts, SSE streaming support, and Pydantic v1/v2 compatibility. External dependencies include `httpx` for HTTP, `pydantic` for models, and optional `aiohttp`, `playwright`, or `pydoll` for advanced browser control. Key entry points are the `Stagehand` and `AsyncStagehand` classes, which can be instantiated with an API key and base URL. Sessions are started via `client.sessions.start()`, then methods like `session.act()`, `session.extract()`, `session.observe()`, and `session.execute()` perform browser automation tasks. The `_custom/session.py` module wraps the generated resource to provide a bound session object with page-aware frame handling. The SDK also ships a `_sea` binary for local execution, packaged via custom Hatch build hooks.
Exosphere is a reliability runtime for AI agents, providing a state-based workflow orchestration platform. The system consists of three main components: a Python SDK (`exospherehost`) for defining nodes and execution graphs, a FastAPI state-manager backend that handles state transitions, retries, cron triggers, and secret encryption, and a Next.js dashboard for monitoring and managing workflows. The architecture follows a microservices-like pattern within a monorepo, with the SDK acting as a client library and the state-manager as the central orchestration engine. Key capabilities include dynamic graph execution, infinite parallel fanout/unite patterns, configurable retry policies (exponential, linear, fixed with jitter), signal-based state pruning and re-queuing, and a graph-level key-value store. The system relies on MongoDB for persistence and uses Beanie ODM. The codebase demonstrates solid engineering practices with well-structured controllers, middleware for request ID propagation and exception handling, and a comprehensive test suite, though some tests are incomplete or skipped. The project is in early beta (v0.0.3b1) with active development and monthly releases.