SuperAPI AI Directory
Partner with us中文AI newsFavorites

AI agents

2,963 results · Page 37 of 83

Bavior

AI agents

Bavior is an AI agent tool that increases a brand's visibility in AI answers through community content such as Reddit. It scans how engines including ChatGPT, Gemini, Perplexity, and Google AI Overviews evaluate a user's product, identifies the Reddit discussion threads these answers rely on, and automatically plans, writes, and publishes posts and replies to try to earn citations by AI. The tool provides a baseline scan starting free, supports approval mode and automatic mode, lets users choose whether to publish with Bavior's established accounts or their own accounts, and includes a record of each action.

Read more

beebot

AI agents

BeeBot is an autonomous AI assistant project, positioned as a personal "worker bee", aimed at autonomously executing a variety of practical tasks. The project selects tools through AutoPack and supports acquiring more tools during task execution; it has built-in persistence, an e2b-compliant REST API, a websocket server for publishing all events, a replaceable filesystem simulation (in-memory, disk, or database storage), as well as dynamically adjusting history during task execution, and optional Helicone caching. Development is currently paused because the author believes that current (late 2023) LLMs are not yet sufficient for general-purpose autonomous AI; the author lists the conditions under which the project might be restarted. Installation requires cloning the repository and using poetry to install dependencies. Persistence is required; SQLite is officially supported, but using Postgres via Docker is recommended. Windows is not officially supported. The project provides a CLI, server startup commands, example API calls for creating and executing tasks, and a websocket connection endpoint; the Web UI is coming soon.

Read more

Bennu AI

AI agents

Bennu AI is an autonomous AI agent that claims to handle operations, deploy code, fix bugs, and keep services running continuously. It offers zero-downtime monitoring, CI/CD automation, security scanning, one-sentence natural language deployment, and other features, and supports integrations with tools such as GitHub, Docker, Prometheus, Kubernetes, and Terraform. Users can describe goals in plain English, and the agent will read configurations, run checks, edit code, update infrastructure, and report results.

Read more

bernstein

AI agents

Bernstein is an open-source governance layer for AI agents that adopts a "policy as code" approach, letting users define declaratively who can do what, which operations require approval, and what content must be recorded, with Bernstein responsible for enforcement and generating verifiable, replayable records. It uses a deterministic scheduler (no model is introduced into the coordination layer), runs agents in parallel in pure Python, gates outputs, and records every step so that run results can be verified after the fact offline using only artifacts. It supports more than 50 CLI coding agents such as Claude Code, Codex, and Gemini CLI. Deliverables can be code diffs, research reports, datasets, or audit evidence packages, and it provides air-gapped installation configurations. The project uses the Apache-2.0 license, is free to use, is currently in beta status, and is maintained by an individual and under active development.

Read more

best-of-Agent-Harnesses

AI agents

best-of-Agent-Harnesses is an AI agent tool aggregation project on GitHub that ranks and displays 167 AI agent harnesses (agent runtime frameworks/hosting tools), along with templates, playbooks, MCP-related content, and learning resources. According to the project description, the list is rescored weekly, making it easy to track ranking changes for each tool. This information all comes from the project's official introduction; the remaining specific item details, scoring methods, usage methods, etc. are not yet provided.

Read more

blinky

AI agents

Blinky is an open-source, VSCode-oriented AI debugging agent that uses an LLM to help locate and fix backend code errors. Inspired by SWE-agent, it combines the VSCode API, the Language Server Protocol (LSP), and print-statement debugging to locate and address bugs in real backend systems. Users need to provide an OpenAI API key on first run, and the key is stored in the local VSCode configuration. Before use, install the Blinky extension from the VSCode Marketplace, click the ghost icon in the sidebar to enter the chat interface, describe the bug to fix and optionally fill in reproduction steps, then start the agent; feedback can be provided at any time during the run, and after it finishes, the changes made by the agent can be accepted or rejected. This repository uses the MIT open-source license, mainly targets backend systems, and plans to expand to other development stacks in the future.

Read more

bmux

AI agents

bmux is a Chromium browser for macOS, Windows and Linux, focused on multi-pane split screen, session management and agent control. It can split web pages into multiple panes, allowing users and AI agents to visibly operate the browser at the same time; it also supports controlling panes from the command line, assigning independent profiles to different agents, simulating mobile device browsing, and setting proxy connections per profile, making it suitable for scenarios requiring multitasking browsing, automated scraping, or letting agents participate in web page operations.

Read more