SuperAPI AI Directory
Partner with us中文AI newsFavorites

AI agents

2,963 results · Page 43 of 83

crewai

AI agents

CrewAI is an open-source Python framework for orchestrating role-playing autonomous AI agents, enabling multiple agents to collaborate on complex tasks through collaborative intelligence. It provides two complementary approaches: Crews and Flows: Crews is aimed at autonomous, collaborative teams of role-based agents, while Flows is aimed at event-driven, precise workflow control. The framework requires Python >=3.10 and <3.14, uses UV to manage dependencies, and can create and run projects via the crewai CLI. The official offering also includes the commercial CrewAI AMP Suite, which includes managed deployment, observability, governance, security, and enterprise support capabilities.

Read more

ctxlint

AI agents

ctxlint is a code linter designed specifically for AI agent context files. It can read context files such as CLAUDE.md, .cursorrules, AGENTS.md, .mcp.json, and cross-check them against the actual codebase, uncovering issues such as broken paths, incorrect commands, outdated references, contradictions between files, and hardcoded secrets in MCP configurations. The tool runs as a command line, can be used via npx without installation, and also supports local project installation. It can be integrated into CI, Git hooks, or run as an MCP server, offering features such as automatic fixes, watch mode, and token consumption statistics, and supports 16 AI tools and 8 MCP clients.

Read more

Cykel

AI agents

Cykel is a digital employee platform that provides AI agents for recruiting, sales, and research scenarios. Its first digital employee is named Lucy, positioned as an AI recruiting partner that can automate talent sourcing, screening, and administrative tasks, helping talent teams shorten the hiring cycle. The platform also provides an AI sales agent that automates lead generation, outreach, and follow-up to accelerate the sales cycle; and an AI business analyst that researches and writes professional-grade company reports for investment, consulting, and advisory teams. All digital employees are claimed to work 24/7.

Read more

DAGent

AI agents

DAGent is a Python library with a clear design orientation, used to quickly build AI agents through a directed acyclic graph (DAG) structure, without additional overhead. Its core idea is to set existing Python functions as nodes in the graph, and have a large language model infer which function should be executed through "decision nodes," thereby implementing agent behavior. Tools are just ordinary functions, and the LLM can call them as tools. The library supports installation (pip install dagent or rye add dagent), and requires the system to provide the required API key, using OPENAI_API_KEY by default. DAGent supports using different LLM models for reasoning and tool description generation, and the model can be specified through call_llm, call_llm_tool, or when compiling a DecisionNode, for example groq/llama3-70b-8192. In use, it is recommended to write docstrings and type annotations for functions to help the LLM understand. The .compile() method automatically generates and saves tool descriptions to the Tool_JSON folder, which can be manually edited to improve reliability. If data is passed between functions, the prev_output parameter must be included in the function signature; if there is a problem with extra parameters, **kwargs can be added. Parameters can be overridden by the user at runtime. This library is suitable for developers who want to quickly build agent workflows by reusing existing Python code.

Read more

Dash

AI agents

Dash is an AI teammate for team work scenarios, usable directly in Slack and Microsoft Teams channels or DMs. It completes tasks through conversation, such as querying HubSpot and posting sales pipeline updates, auditing Google Ads and giving scale-up or pause recommendations, categorizing Zendesk tickets and drafting replies, and fixing code issues and submitting PRs. Dash can connect to up to 3,000+ tools; once configured, the whole team shares tools, skills, and context, with no need to learn a new app. For risky operations (sending, posting, writing, spending), it requests approval first; non-mutating reads show results and clickable sources. The official description states that its workspaces are isolated from one another, it does not use data to train models, and it has SOC 2 Type II and ISO/IEC 27001 certifications.

Read more

Daski

AI agents

Daski is a curated procurement marketplace for AI agents, used to discover, evaluate, purchase, and manage real services such as domains, email, and company registration. Its core feature is that it allows AI agents to autonomously purchase real services: agent-initiated, machine-orchestrated, paid via USDC and settled on the Base chain. Information shows that Daski positions itself as a coordination layer, discovery and payment quotes flow through Gateway, settlement is completed on-chain, and buyers and sellers communicate directly about everything else. Users can copy a prompt to an agent, enabling it to purchase services on the Daski marketplace. Currently, official information does not specify concrete pricing, fee model, service coverage, or usage thresholds; this information is not yet available.

Read more