Business agents

A tiny, dependency-free tool belt for LLMs: token counting, cost estimation, retries, prompt templates, and text chunking.

pprof for your LLM context - see where every token and dollar goes.

Local-first context planning and compression for LLM applications, RAG systems, and agents.

AutoGen v0.4 MemoryStore adapter for the Mnemosyne memory substrate. Implements the MemoryStore protocol with per-agent container scoping, hybrid BM25+vector recall, async support for high-throughput multi-agent runs.

Ontology-guided knowledge graphs and graph-grounded RAG for document intelligence

Multi-provider AI agent CLI supporting Ollama, Bedrock, OpenRouter, OpenAI, Meta, Gemini, xAI, Cloudflare, and Anthropic

Cross-project LLM agent runtime for Dragonflow — loads agent rows from a per-project `agents` Supabase table and invokes Anthropic / Google / OpenAI models against the v0.1+ invocation contract.

E2E parity validator for site migrations. Compares prod vs cand and reports UI, functional, SEO, visual, and Web Vitals deltas with an LLM-ranked HTML report.

Vigil — a defensive-cyber terminal agent for Computer Network Defense. Asset discovery, vulnerability assessment, hardening, detection engineering, threat hunting, incident response — on the infrastructure you defend.

A headless, multi-provider AI agent harness. One install gives you the `cortex` CLI and HTTP server: npm i -g nexus-cortex.

Sanctum Runtime SDK — the runtime trust boundary for autonomous systems. Signed action tokens, blast-radius scoring, source-trust classification (indirect-prompt-injection defense), dual-approver + auto-escalate, policy replay, SOC2 + NIST AI RMF evidence

AWS Bedrock Agents governance middleware for Aegis ACP (maintenance-only — Anthropic SDK is the active hero)

One-line spending control for AI agents. @budget(max_usd=10.00) decorator for LLM APIs and x402 payments.

First-Class Hermes Tool Agent — Offline ReAct agent with dual memory (episodic + semantic), skill learning, and daemon mode

LangChain integration for rezunate-llm-sdk

Shrink LLM context windows by removing noise, redundancy, and long-tail detail — without losing signal.

MothRAG: adaptive multi-arm subset routing for multi-hop retrieval-augmented generation — high-level Python API with auto-defaults.

Token-efficient prompting and agentic workflows for Claude

A small, framework-agnostic LLM-in-a-loop primitive: zod-validated tool registry, 5 independent stop conditions (steps / tool-calls / wallclock / tokens / cost), and trace/budget output. SDK-agnostic via a single AgentStepFn seam.

OpenAI / Anthropic tool-calling adapter for cdr-kit — JSON-Schema tools[] + a dispatch router.

Heal broken Vercel AI SDK message arrays before they hit the provider. Fixes orphaned tool calls, missing reasoning signatures, invalid tool names, and other provider rejections.

一个用于测试的 Mock LLM 服务器，支持 OpenAI、Anthropic 和 Responses API 格式

ESLint for your LLM prompts — catch prompt-injection risks, missing token limits, and prompt anti-patterns before you ship.

Paybyrd integration for a claude AI agent

A protocol and reference implementation for user-customizable web app interfaces. Apps declare schemas, actions, and primitive components. Users (via an LLM agent) compose those primitives into custom views, parameterized by a theme token vocabulary. The

Logging and tracing layer for AgentsKit agents.

Coinbase AgentKit ActionProvider for trading Thetanuts options on Base — for autonomous backend agents that own their own wallet (CDP, viem, Privy, server wallets).

Unified LLM API with automatic model discovery and provider configuration

Agente de navegador autónomo con IA (Claude). Abre Chrome o Firefox reales, lee la página y la opera con una sola instrucción en lenguaje natural. Incluye truco CDP anti-Cloudflare.

Lightweight Python client for the Agentic Trading Lab REST API (LLM-powered trading agents, backtests, paper trading).

FreshRSS API + MCP Server + A2A Server

A pre-spend GATE + learning advisor for LLM API cost: caps every call, prices from a verified table, and learns the cheapest config that keeps quality.

MemGPT memory architecture as a Python sidecar for the openclaw-memgpt plugin. Includes Anthropic/Claude compatibility patches against the original MemGPT (Apache 2.0).

Qode — LLM provider abstraction layer (Anthropic, OpenAI, Google, Ollama, etc.)

Access AI models, email services, and project thumbnail capture through vly.ai and freebuff.com without managing multiple API keys or accounts

Papaya AI agent trace capture SDK for production apps.

A small-but-complete agent harness: ReAct loop, tool injection, three-tier memory, restricted-expression panels. Zero heavy deps.

The language-agnostic governance-gate sidecar — govern a non-MCP agent runtime (LangGraph, CrewAI, AutoGen) by remoting each native tool call through the Action Kernel over a thin NDJSON-RPC seam. The reusable spine of Lodestar's runtime-adapter epic.

AgentMetrics observability integration for Claude Managed Agents

Memory-backed autonomous agent with a Docker membrane, a Slack/Discord cockpit, and capability graduation.

Govern an AutoGen agent's native tool calls with Lodestar — the thin native hook that remotes each tool call through the Lodestar Action Kernel over NDJSON-RPC (ADR-0027).

Govern a CrewAI crew's native tool calls with Lodestar — the thin native hook that remotes each tool call through the Lodestar Action Kernel over NDJSON-RPC (ADR-0026).

Govern a LangGraph agent's native tool calls with Lodestar — the thin native hook that remotes each tool call through the Lodestar Action Kernel over NDJSON-RPC (ADR-0024).

AI Agent Skills management tool - Install and manage Claude/AI agent skills

Pi extension that keeps noisy tool output out of your context window. Automatically rewrites shell commands through Hypa for local, deterministic compression, context-aware file tools, and recoverable evidence.

Shared wire types for the Threadbase multi-agent pipeline (tb-multi-agent ↔ tb-streamer).

Agent-friendly Markdown-to-PDF CLI with HTML preview and JSON diagnostics

Define agents (tools + system prompt) in YAML and run them as an orchestrated pipeline.

Que Claude genere interfaces de usuario con A2UI en sus respuestas.

CrewAI toolkit for Ceki — real Chrome sessions for AI agents with persistent profiles across runs

LangChain tool for Ceki — real Chrome sessions for AI agents

The billing layer for AI agents — self-hosted, Stripe-integrated usage metering and prepaid-credit billing for LLM/agent spend.

Sales Prospecting review: compare features, pricing, use cases, access model, and alternatives for this Sales agent in 2026. AI prospecting + LinkedIn outreach for B2B teams. Find buyers, automate follow-ups, track replies, a

Cyrene: the AI agent that always loves you — a self-improving, open-source AI agent written in Rust. This package installs the prebuilt native binary for your platform.

LangChain (JS) tools for Telaro. Drop these into any LangChain agent — OpenAI, Anthropic, Gemini — to query trust scores, gate capital delegation, and (optionally) auto-record actions on Solana.

Runtime AI agent security SDK for MastGuard — prompt injection detection, scope enforcement, tamper-evident audit logging, and policy-based guardrails.

The autonomous-agent harness that actually finishes long tasks — because it plans first, self-corrects every step, and survives crashes. The whole run lives in one serializable RunContext, checkpointed after every step, so it resumes exactly where it died

Agentic memory for Pi — verbatim session capture + grep retrieval

Lightweight multi-agent orchestration for LLMs. Works with OpenAI and Anthropic. Pipeline → Router → Agent with priority short-circuits, session memory, and tool use.

A friendly, batteries-included toolkit for building AI agents in Python.

Run a prompt through a LangChain (system + human) chain over Gemini on Vertex AI; optional LangSmith tracing.

A lightweight, zero-dependency, provider-agnostic LLM agent with tool-calling support for Anthropic, OpenAI, Gemini, and Ollama.

The FastAPI of AI Agents — build production-ready AI agents, teams, RAG, and workflows from one package.

agentto connector — links your locally-installed AI agent CLIs through link.agentto.net to Agent to Dashboard (dash.agentto.net). Run `agentto` for a TUI.

Anthropic Claude phase executor for Coro — wraps the Claude Agent SDK and exposes it through the PhaseExecutorRuntime contract.

LangChain.js tools for Agent402.Tools (the applied layer of Agentic Finance): turn its catalog - 500+ strong: 400+ pay-per-call tools (x402 or MPP) + 70+ skill packs (browser, web search, OCR, PDFs, memory) - into LangChain tools via four meta-tools (find

LangChain.js BaseChatMemory implementation backed by Mnemosyne. Persistent, searchable, container-scoped chat history with hybrid BM25+vector recall. Replaces LangChain's in-process BufferMemory and ZepMemory. Per-conversation thread tracking, automatic r

Track AI agent costs, tokens, runtime and spending. Prevent runaway OpenAI, Anthropic, LangGraph and OpenRouter agents from exceeding budget.

关系感知 Skill — 多 Agent 系统的社会层

Lightweight SDK for tracking AI agent costs and reliability with AgentAxon

A reactive declarative UI and threading utility library for exteraGram / Telegram plugins.

Anthropic component(s) as a standalone Langflow Extension Bundle.

The agent framework where the model never holds the trigger — every consequential action clears your policy first, waits for a human when it matters, and lands on a record you can verify. Build on it, or put it around the agent you already have.

AI Setu client — drop-in OpenAI SDK with workspace credit accounting and multi-provider routing.

The AnyRouter provider for the Vercel AI SDK — call every AnyRouter model through `createAnyRouter`.

Track AI agent costs, tokens, runtime and spending. Prevent runaway OpenAI, Anthropic, LangGraph and OpenRouter agents from exceeding budget.

The model-independent trust layer for AI agents: verifiable identity, model provenance, tamper-evident SHA-256 audit chains, policy enforcement, and outcome verification. Framework-agnostic.

A minimal, dependency-light kernel for a general-purpose AI agent — runs anywhere, brings its own brain.

Argus — AI threat-modeling agent (multi-agent, skills)

Axemere AI Gateway Anthropic SDK wrapper — drop-in Anthropic/AsyncAnthropic subclasses

One-line spending control for AI agents. @budget(max_usd=10.00) decorator for LLM APIs and x402 payments.

Tiered, token-budget-aware memory for any LLM agent

Provider-agnostic token profiler for LLM agent loops. See exactly where your context tokens go — and how many you waste re-sending the same thing.

二郎神 - 全知全能的AI投资智能体

Universal AI agent runtime — portable .agent files run identically in cloud, desktop, CLI, and as an MCP server. Multi-LLM (Anthropic, OpenAI, Gemini), pluggable adapters, MCP client + server, streaming, abort, pause/resume.

General-purpose agent with transport abstraction, state management, and attachment support

Agent evaluation and benchmarking for AgentsKit.

Paybond Kit for TypeScript: agent spend governance for paid tool calls with spend authorization, evidence receipts, refunds, disputes, hosted Gateway sessions, and settlement.

Anthropic provider tokens and model factory wiring for Nest LangChain.

Unbrowse as native LangChain JS tools. Three tools (resolve, execute, search) shaped exactly like LangChain's StructuredTool/DynamicStructuredTool — { name, description, schema (JSON Schema), invoke, call }. No zod dependency. Pass your own @langchain/cor

Custom Provider CLI for managing model providers

efferent — an agent runtime on Effect.ts + Bun. One CLI to run agents: an interactive TUI, headless (print/json/rpc), or a persistent per-workspace daemon that thin clients attach to. Multi-provider (Anthropic/OpenAI/Gemini), persistent sub-agent context

Multi-tier AI orchestration CLI — T1/T2/T3 hierarchical agent system

AI assistant plugin for Payload CMS with provider selection, CMS mentions, and signed action proposals.

Production utilities for Vercel AI SDK — streaming helpers, cost tracking, provider fallback, caching, and structured output validation

Orchestrated multi-agent pipeline for Jira tickets using Cursor SDK

AI Agent Framework for Salesforce — programmatic context, tools, plugins, and agent lifecycle management.

Production-ready AI agent framework with persistent memory, self-reflection, and adaptive planning

Backend middleware for the Threadplane client-tools capability. The /langgraph entrypoint targets LangGraph.js.

A testing tool for exploring congestion and fault propagation in asynchronous systems