App Builders

Trickle

Trickle introduces an agentic canvas where teams and AI collaborate in real time to build production-ready apps and websites without starting from scratch.

• Visual, node-based canvas for designing data flows and UI

• Built-in AI agents for code generation and content creation

• Drag-and-drop components with instant preview

• Export to React, Vue, or static HTML/CSS

• One-click deploy to Vercel, Netlify, or custom domains

URL: https://trickle.so/

Blink.new

Blink.new accelerates software prototyping by scaffolding full-stack templates and letting AI refine your code, so you can launch MVPs in minutes.

• Prebuilt templates for web apps, APIs, and microservices

• AI-powered code suggestions and completions

• GitHub integration for seamless commits and pull requests

• Real-time collaboration and live preview

• Custom domain binding and hosting setup

URL: https://blink.new/

Open Lovable

Open Lovable is the community-driven, open-source remake of the Lovable platform for building AI-powered web applications with minimal code.

• Drag-and-drop UI editor for forms, dashboards, and chatbots

• Plugin ecosystem for adding third-party integrations

• Built-in state management and data fetching

• Themeable components and layout library

• MIT license with full source code on GitHub

URL: https://github.com/mendableai/open-lovable

AI Agents

Macaron

Macaron is your highly personal AI companion that learns from every interaction, adapts to your preferences, and helps automate day-to-day tasks.

• Personalized memory layer retaining user context over time

• Multi-modal chat interface supporting voice and text

• Customizable “personality profiles” for different use cases

• API hooks to integrate with your calendars, emails, and apps

• Freemium tier with priority upgrades for power users

URL: https://macaron.im/

Parlant 3.0

Parlant 3.0 offers an open-source engine for modeling agent behaviors, enabling you to orchestrate complex multi-agent workflows and simulations.

• Behavior trees and finite-state machines for agent logic

• Plugin architecture for custom sensors and actuators

• Support for synchronous and asynchronous agent coordination

• Visualization tools for monitoring agent states and interactions

• Dual licensing: open-source core with premium extensions

URL: https://github.com/emcie-co/parlant

UI-TARS-desktop

UI-TARS-desktop brings a GUI-driven agent stack to your machine, blending terminal power with vision-based AI assistants in a unified interface.

• Cross-platform desktop app (Windows, macOS, Linux)

• Integrated computer vision for drag-and-drop UI automation

• “Agent cards” that bundle workflows, scripts, and models

• CLI mirror for headless server deployments

• Freemium model with pro agent packs

URL: https://github.com/bytedance/UI-TARS-desktop

AI Models & AI Development

Letta AI

Letta AI provides an open-source memory layer that supercharges your LLM and agent applications with lifelong learning and retrieval-based reasoning.

• Vector-based memory store with similarity search

• RAG (Retrieval-Augmented Generation) connectors for major LLMs

• Automatic memory summarization and pruning

• REST API and SDKs in Python and JavaScript

• Community edition under MIT license

URL: https://www.letta.com/

MemU

MemU adds persistent, context-aware memory to any AI agent, ensuring consistent conversational experiences and deeper reasoning across sessions.

• Graph-based knowledge representation

• Session replay and context stitching

• Plugin system for custom memory schemas

• Lightweight embedding store for on-device use

• Freemium plan with enterprise options

URL: https://memu.pro/

Jan V1

Jan V1 is an open-source, 4 B-parameter LLM fine-tuned for web search queries—an on-premise alternative to proprietary search assistants.

• Optimized tokenization for search intent recognition

• Small footprint for low-cost GPU inference

• Built-in index retriever for local document corpora

• Zero-shot Q&A and summarization modes

• Available on Hugging Face under Apache 2.0

URL: https://huggingface.co/janhq/Jan-v1-4B

Dots.OCR

Dots.OCR is a 1.7 B-parameter vision-language model delivering state-of-the-art performance on multilingual document parsing and layout analysis.

• High-accuracy text extraction from scanned docs and PDFs

• Table structure recognition and CSV export

• Support for complex scripts (Chinese, Arabic, Devanagari)

• End-to-end PyTorch implementation with pretrained weights

• Open-source under MIT license

URL: https://github.com/rednote-hilab/dots.ocr

Creative AI

Higgsfield Product to Video

Flowstep

Flowstep leverages AI to generate UX/UI designs on demand—draft wireframes, style guides, and interactive prototypes in seconds.

• Prompt-driven design generation with theme presets

• Export to Figma, Sketch, or Adobe XD formats

• Component library with accessible, responsive elements

• Style-consistency checker and design tokens exporter

• Freemium access with team collaboration features

URL: https://flowstep.ai/

Qwen Image Editing

Qwen Image Editing offers an AI model specialized in high-fidelity image transformations, from background removal to style transfer.

• Text-prompted edits: remove, replace, or stylize elements

• Inpainting and outpainting for seamless extensions

• Built-in upscaling up to 4× without visible artifacts

• REST API with batch processing support

• Freemium tier for casual users

URL: https://chat.qwen.ai/?inputFeature=image_edit

Marketing & Productivity

Go High Level

Go High Level is an all-in-one growth platform that unifies CRM, marketing automation, and analytics to power campaigns end to end.

• Pipeline management and lead scoring

• Automated SMS, email, and voicemail sequences

• Drag-and-drop funnel and landing page builder

• Two-way conversations with inbound leads

• White-label options for agencies

URL: https://www.gohighlevel.com/

Kuse AI

Kuse AI is an AI-driven canvas where you can feed files, links, and videos into a chat interface and surface actionable insights and workflows.

• Multi-modal ingestion: PDF, DOCX, URLs, media

• Context-aware query parser with SQL-style filters

• Workflow builder to chain AI tasks (summarize→analyze→assign)

• Team sharing and real-time collaboration

• Freemium access with per-seat upgrades

URL: https://app.kuse.ai/