The Ollama Python library provides the easiest way to integrate Python 3.8+ projects with Ollama. The Ollama Python library's API is designed around the Ollama REST API.
Open source monitoring for electricity, solar, storage, heat pumps and electric vehicle charging. A versatile and expandable system of sensors and integrations built on the Raspberry Pi and Arduino platforms. Expandable. Extensible. Self hosted.
μlogger is an android application for continuous logging of location coordinates, designed to record hiking, biking tracks and other outdoor activities. Application works in background. Track points are saved automatically at chosen intervals or manually and may be uploaded to dedicated server in real time. This client works with μlogger web server. Together they make a complete self owned and controlled client–server solution.
Meant to be simple and small (μ). Low memory and battery impact. Focuses on privacy, doesn't use Google Play services, logs to self-owned server. Uses GPS or network based location data. Synchronizes location with web server in real time; in case of problems keeps retrying. If you're offline it'll queue updates until you're back online. Allows adding waypoints with attached images and comments (required μlogger server version 1.0+ for synchronization).
In the Google Play store.
This is a web application for real-time collection of geolocation data, tracks viewing and management. Together with a dedicated μlogger mobile client it may be used as a complete self hosted server–client solution for logging and monitoring users' geolocation.
Requires PHP v5.5 (with ctype, json, pdo, session, simplexml, and xmlwriter enabled), a back-end database (MySQL, Postgres, or SQLite), and a mobile app that sends waypoints to the server.
Private. If you set it up nobody can see it unless you give them an account, and you can't just create an account in an ad-hoc fashion. Tracking software requires you to log in so you can't just throw coordinates at it. Doesn't have a REST API, either (though I wish it did). Runs nicely in shared hosting.
Personal link database, link aggregator, with RSS functionality. Has a search engine for your links database. Basic data analysis features - analyze link rot, how many a page is cited by other sources, analyze link domains, etc. Tags and comments on links. Multiple user accounts are possible. Data can be exported for storage and importation elsewhere. Support for 'spaces'. You can define own spaces like 'music', 'videos', 'movies', etc. Implemented as separate Django apps. Keyword entry analysis to find trends.
smolagents is a library that enables you to run powerful agents in a few lines of code. The logic for agents fits in about 1,000 lines of code. Our CodeAgent writes its actions in code (as opposed to "agents being used to write code"). To make it secure, we support executing in sandboxed environments via E2B or via Docker. Supports any LLM. It can be a local transformers or ollama model, one of many providers on the Hub, or any model from OpenAI, Anthropic and many others via our LiteLLM integration. Agents support text, vision, video, even audio inputs!
Nerve is an ADK (Agent Development Kit) designed to be a simple yet powerful platform for creating and executing LLM-based agents. Agents are simple YAML files that can use a set of built-in tools such as a bash shell, file system primitives and other things (like APIs).
(archived) https://www.evilsocket.net/2025/03/13/How-To-Write-An-Agent/
A commandline utility to search text in PDF files. Tries to be compatible with GNU Grep, where it makes sense. Many of your favorite grep options are supported (such as -r, -i, -n or -c).
Git: https://gitlab.com/pdfgrep/pdfgrep
I wonder if I can plug this into SearxNG.
AutoKitteh is a developer platform for workflow automation and orchestration. It is an easy-to-use, code-based alternative to no/low-code platforms (such as Zapier, Workato, Make.com, n8n) with unlimited flexibility. You write in vanilla Python, we make it durable. Once installed, AutoKitteh is a scalable "serverless" platform (with batteries included) for DevOps, FinOps, MLOps, SOAR, productivity tasks, critical backend business processes, and more.
Provides interfaces for building projects (workflows), deploying them, triggering the code with webhooks or schedulers, executing the code as durable workflows, and managing these workflows. All services are available via gRPC / HTTP. Has a CLI as well.
An implementation of Eno and Schmidt's Oblique Strategies. Done entirely in HTML5, with just enough Javascript to pick randomly. Might even work as a PWA.
Beszel is a lightweight server monitoring platform that includes Docker statistics, historical data, and alert functions. It has a friendly web interface, simple configuration, and is ready to use out of the box. It supports automatic backup, multi-user, OAuth authentication, and API access.
Smaller and less resource-intensive than leading solutions. Easy setup, no need for public internet exposure. Configurable alerts for CPU, memory, disk, bandwidth, temperature, and status. Users manage their own systems. Admins can share systems across users. Supports many OAuth2 providers. Password auth can be disabled. Save and restore data from disk or S3-compatible storage.
Consists of a hub built on PocketBase that provides a dashboard for viewing and managing connected systems and an agent that runs on each system you want to monitor, creating a minimal SSH server to communicate system metrics to the hub.
Seems to be designed with system monitoring in mind, and as such isn't really that flexible.
A Self-hosted .ics feed aggregator that publishes an anonymized public schedule to a GitHub repository. A light weight, ultra simplified, copyleft alternative to cal.com. Looks like it outputs Markdown, which would be ideal for a repo at Github. Anonymizes event details (can you turn this off?) Publishes formatted schedule to public repo (can we change this to private?) Automatic syncing via cron.
Ace is a standalone code editor written in JavaScript. Our goal is to create a browser based editor that matches and extends the features, usability and performance of existing native editors such as TextMate, Vim or Eclipse. It can be easily embedded in any web page or JavaScript application. Ace is developed as the primary editor for Cloud9 IDE and the successor of the Mozilla Skywriter (Bespin) Project.
Syntax highlighting for over 120 languages. Over 20 themes. Automatic indent and outdent. An optional command line. Handles huge documents (at last check, 4,000,000 lines is the upper limit). Fully customizable key bindings including vim and Emacs modes. Search and replace with regular expressions. Highlight matching parentheses. Toggle between soft tabs and real tabs. Displays hidden characters. Drag and drop text using the mouse. Line wrapping. Code folding. Multiple cursors and selections. Live syntax checker (currently only JavaScript/CoffeeScript/CSS/XQuery). Cut, copy, and paste functionality.
You do not generally need to build ACE. The ace-builds repository endeavours to maintain the latest build, and you can just copy one of src/* subdirectories somewhere into your project. I'd probably feel safer downloading a release, to be honest.
Marreta is a tool that breaks access barriers and elements that hinder reading. Automatically cleans and corrects URLs. Removes annoying tracking parameters. Forces HTTPS to keep everything secure. Changes the user agent of the request to avoid blocking. Leaves HTML clean and optimized. Fixes relative URLs on its own. Allows you to add your own styles and scripts. Removes unwanted elements. Caches all the things. Blocks domains you don't want. Allows configuring headers and cookies however you want.
Written in PHP. No database required.
A browser-based WARC file viewer. No data is uploaded anywhere and no information is collected. All content rendered stays directly in your browser. When loading an archive from Google Drive, the site may ask for account authorization to download the specified file only.
Github: https://github.com/webrecorder/replayweb.page
ReplayWeb.page provides a static site generated with MkDocs, an npm package/library, and an Electron app all in this repo.
This repository contains the 'frontend' UI for the replay system, while the 'backend' is provided via a service worker implementation found at: https://github.com/webrecorder/wabac.js. (Of course, both frontend and backend actually run in the browser).
The frontend is loaded from ui.js, while the backend service/web worker is loaded from sw.js.
Use py2mappr to render fully interactive networks from simple csv files of node and links. Use tag2network to build a similarity network from a list of nodes with associated tags (e.g. documents with keywords). Nodes can host images, video, Twitter and Instagram feeds, and music streams.
Easily style the appearance of nodes, links, and node labels. Reveal patterns in network structure with scatterplot and geospatial layouts. Save custom views and layouts with seamless animated transitions between them. Share a fully interactive network by publishing openmappr files to any static website.
Code-free data exploration with dynamic filters and search tools to query nodes by any combination of node characteristics. Effortlessly zoom in to node details and zoom out to see them in the broader network context. See summaries of node characteristics - both for the entire network and selected groups of nodes.
You'll have to explore a bit to find the tools' links to Github. It could be better organized.
FlareSolverr starts a proxy server, and it waits for user requests in an idle state using few resources. When some request arrives, it uses nodriver or Selenium with the undetected-chromedriver to create a web browser (Chrome). It opens the URL with user parameters and waits until the Cloudflare challenge is solved (or timeout). The HTML code and the cookies are sent back to the user, and those cookies can be used to bypass Cloudflare using other HTTP clients.
NOTE: Web browsers consume a lot of memory. If you are running FlareSolverr on a machine without a lot of RAM, do not make many requests at once. With each request a new browser is launched.
It is also possible to use a permanent session. However, if you use sessions, you should make sure to close them as soon as you are done using them.
A pixel-perfect web-based MS Paint remake and more. Ah yes, good old Paint. Not the one with the ribbons or the new skeuomorphic one with the interface that can take up nearly half the screen. (And not the even newer Paint 3D.) Recreates every tool and menu of MS Paint, and even little-known features, to a high degree of fidelity. It supports themes, additional file types, and accessibility features like a Dwell Clicker and Speech Recognition. Claims to be mobile friendly.
You can create links that will open an image from the Web in JS Paint. Rudimentary multi-user collaboration support. It isn't seamless; actions by other users interrupt what you're doing, and visa versa. Sessions are not private, and you may lose your work at any time.
jsPaint can be installed as a Progressive Web App (PWA), although it doesn't work offline yet. Look for the install prompt in the address bar.
A lightweight and cross platform QR Code and Bar code scanning library for the web. Support scanning different types of bar codes and QR codes. Supports different platforms and browsers. Supports scanning with camera as well as local files. Comes with an end to end library with UI as well as a low level library to build your own UI with.
Based on zxing-js.
A Barcode Detection API polyfill that uses ZXing-C++ WebAssembly under the hood.
Supported barcode formats: aztec, code_128, code_39, code_93, codabar, databar, databar_expanded, databar_limited, data_matrix, dx_film_edge, ean_13, ean_8, itf, maxi_code (only generated ones, and no position info), micro_qr_code, pdf417, qr_code, rm_qr_code, upc_a, upc_e, linear_codes and matrix_codes (for convenience).
It's probably easier just to grab it from the JS CDN and store it locally.