SuperAPI AI Directory
Partner with us中文AI newsFavorites

AI agents

2,941 results · Page 39 of 82

ChatHelp

AI agents

ChatHelp is an AI tool and agent aggregation platform for small and medium-sized businesses, office, and learning scenarios, offering a free website AI chat plugin. Its core capabilities include: training a custom customer service chatbot based on website content, supporting documents, product descriptions, PDFs, and other materials, embedding it into a website with one line of code, answering visitor questions in real time, and supporting taking over human conversations at any time and automatically collecting potential customer leads. The platform supports the latest AI models such as OpenAI ChatGPT, Anthropic Claude, and Google Gemini, allows users to bring their own API keys, and covers more than 175 languages. It also provides an AI writing assistant with automatic WordPress publishing, AI avatar/asset/art image generation, Yelp review analysis, and other tools, as well as the ability to create custom AI knowledge bases (supporting secure sharing or public access). It supports integrations with WhatsApp, Instagram, Facebook Messenger, Telegram, Discord, Line, Slack, and others, and can be connected through a WordPress plugin. Supported file formats include pdf, doc, docx, ppt, pptx, xls, xlsx, csv, json, epub, mp3, jpg, jpeg, png, and YouTube transcript text. The materials state that files are hosted in an encrypted cloud, and the AI is trained only on encrypted data fragments.

Read more

chemcrow-public

AI agents

ChemCrow is an open-source package for accurately solving reasoning-intensive chemistry tasks, developed by ur-whitelab and published on GitHub. It is built on Langchain and integrates a variety of chemistry tools, including RDKit, paper-qa, and related chemistry databases such as PubChem and chem-space. Due to API usage restrictions, the repository does not include all the tools described in the ChemCrow paper, so its results may not be fully consistent with the paper. All experiments have been published under ChemCrow runs. Users can try it on HuggingFace. Before use, API keys need to be set in the environment, and the Serp API is optional. By default, ChemCrow relies on the RXN4Chem API for retrosynthesis planning and reaction product prediction, which may be slow and requires an API key; alternatively, these tools can be self-hosted using prebuilt Docker images.

Read more

ciagent

AI agents

CIAgent is an open-source tool for checking whether AI agent evals are reliable. By running the same eval multiple times (--runs N), it exposes the decisions that actually flip beneath a surface of stable overall scores, and attributes each flip to a specific source: agent-variance (the agent's output changes, requiring fixes to prompts, retrieval, or temperature), judge-flake (the output is the same but the LLM judge changes its mind, requiring fixes to the grading criteria or a switch to deterministic checks), and infra-error (the judge API call fails, so retrying is enough). It also provides LLM judge auditing (judge-audit), which re-scores with existing ground truth to compare the judge's consistency with deterministic checks, and outputs a TRUSTWORTHY / NEEDS CALIBRATION / UNRELIABLE verdict. CIAgent advocates prioritizing deterministic code fact checks (expected_in_answer, not_in_answer, regex_match, json_schema), path checks, and cost budgets over expensive LLM judging, and can also mine factual assertions from a knowledge base. It natively supports OpenAI, Anthropic, and LangGraph, and can import trace data exported by OpenTelemetry, Langfuse, LangSmith, and others, and can run directly within pytest.

Read more

Claude Skills Library

AI agents

Claude Skills Library is a curated directory of Claude skills and tools, designed to provide Claude AI agents with real-world data and capabilities. The website brings together agent skills and workflows that can be used with Claude, Claude Code, Codex, and Cursor, supporting browsing by use case, quick downloads, and contributing your own skills via GitHub. The library includes skills covering multiple areas such as SEO, UI/UX design, front-end development, the Notion API, Git commit assistance, and more. It is officially stated that the database contains over 90,000 skills, most of which are free and open source, and it provides one-click installation commands for Claude Code and OpenCode.

Read more

claude-code-pro-pack

AI agents

claude-code-pro-pack is a collection of behavioral rules and skill templates for AI coding agents, suitable for Claude Code, Codex, Cursor, and Hermes Agent. At its core is the CLAUDE.md or AGENTS.md twelve-rule behavior file, which claims to reduce AI coding errors from about 40% to about 3%, with a total size of about 700 tokens, and it takes effect simply by placing it in the project root directory. The pack is built on real agent failure patterns from more than 30 codebases, supplementing agent orchestration issues not covered by the original 4 rules, such as token spirals, silent partial failures, dual-mode contamination, and duplicate function drift. The project uses the MIT license and includes the cc-audit auditing tool and example skill templates.

Read more