DEV Community: Slayer King

I Built DeepSearch: An AI Search Assistant for Public Information

Slayer King — Sun, 25 Oct 2026 12:59:07 +0000

I'm the developer of DeepSearch AI Search Lens, published by my company CYBORGATE LTD (https://cyborgate.uk). Here's the problem it addresses and how it works.

The problem

Researching a public figure, a company or a trending topic usually means opening a dozen tabs: a search engine, an encyclopedia, a video site, news, image search. Then you have to piece it all together. Students, journalists, recruiters and curious readers all spend time on that same repetitive routine. I wanted to bring those results together in one place and make it easy to ask follow-up questions.

How DeepSearch helps

DeepSearch is an AI-powered search assistant for exploring publicly available information from across the web. You can search by name or username, or look up public figures, companies, organisations, scientists, authors, athletes, artists, entrepreneurs, historical figures and trending topics.

Instead of checking many sites one by one, DeepSearch gathers relevant public profiles, images, videos and web results into one clean interface. With the Ask AI feature, you can dig deeper into what you find.

Key features

Search people, organisations and topics by name or username
Combined results: public profiles, images, videos and web results in one place
Ask AI to explore results in more depth
Trending searches and popular topics
Search history so you can pick up where you left off

An honest note on privacy

This matters to me. DeepSearch only shows information that is already publicly available. It doesn't generate or expose private information. It's a research convenience tool, not a way to dig up things people haven't made public. Please use it responsibly and respect other people's privacy.

Like any AI tool, its summaries can be incomplete or wrong. Check important facts against the original sources it links to.

It's currently Android only. There's no iOS version yet.

Who it's for

Students and researchers gathering background on a topic
Journalists and writers doing initial research
Recruiters and business professionals looking up companies and public profiles
Content creators researching trends
Curious learners who want quick, organised answers

Try it

Google Play: https://play.google.com/store/apps/details?id=com.cyborgate.aiassistant.peoplefinder.deepsaerch
App Store: not available yet

I'd value your feedback

I'm still refining DeepSearch, and I'd really like your input:

Are the combined results relevant and well organised?
How useful are the Ask AI answers for your research?
Would an iOS version matter to you?
What responsible-use safeguards would you expect to see?

If you build search or retrieval tools, I'd like to hear how you balance usefulness and privacy. Leave a comment below!

I Built VPN MAX Pro: A Simple One-Tap VPN for Private Browsing

Slayer King — Sun, 25 Oct 2026 12:57:40 +0000

I'm the developer of VPN MAX Pro – Fast Proxy, published by my company CYBORGATE LTD (https://cyborgate.uk). In this post I explain what it does and who it's for.

The problem

Some people want every setting and protocol option in their VPN. Plenty of others want to tap a button, see that they're protected and get on with their day. VPN MAX Pro is for the second group. I built it to keep the experience as simple as possible.

How VPN MAX Pro helps

VPN MAX Pro encrypts your internet traffic and routes it through secure servers, which helps protect your data on Wi-Fi and mobile networks. One tap connects or disconnects you, and you can choose from high-performance server locations worldwide for a stable connection.

On iOS, it uses Apple's native VPN framework to set up encrypted tunnels. I chose this because it's well integrated with the operating system.

Key features

One-tap connect and disconnect
Worldwide server locations
Encrypted tunnels using Apple's native VPN framework on iOS
Privacy-focused design that doesn't monitor or record your browsing activity
Session statistics: server location, connection duration and data usage
In-app speed check

How it differs from my other VPN app

I also build V Shield, which highlights WireGuard and Shadowsocks protocol support. VPN MAX Pro focuses on simplicity, native integration and a built-in speed check. If you're choosing between them, V Shield is the one for protocol options, and VPN MAX Pro is the one for keeping things simple.

An honest note

A VPN protects your traffic between your device and the VPN server. That's most valuable on untrusted networks like public Wi-Fi. It won't make you anonymous or protect you from phishing or weak passwords, so please keep good security habits too.

Some features require a subscription, and the plans are shown in the app. I'd rather you know that before installing.

Who it's for

People who want straightforward VPN protection without configuration
Anyone who often uses café, airport or hotel Wi-Fi
Travellers protecting their traffic on the go
Users who like to check their connection speed

Try it

App Store: https://apps.apple.com/gb/app/vpn-max-pro-fast-proxy/id6759932250
Google Play: https://play.google.com/store/apps/details?id=com.cyborgate.vpnmaxpro

Help me improve it

I'd like to hear honest feedback:

How does connection speed compare to other VPNs you've used?
Are there server locations you'd like to see?
Is the app as simple as it should be, or did anything get in your way?

If you've built network or VPN tooling on mobile, I'd also like to swap notes. Comments are open below!

I Built V Shield: A One-Tap VPN with WireGuard and Shadowsocks

Slayer King — Sun, 25 Oct 2026 12:56:28 +0000

I'm the developer of VPN – Super Fast Ultra Proxy (V Shield), published by my company CYBORGATE LTD (https://cyborgate.uk). Here's an honest overview of what it does.

The problem

Public Wi-Fi in cafés, airports and hotels is convenient, but you usually don't know who else is on the network. Many people know they should use a VPN, but lots of VPN apps are cluttered, slow or confusing. I wanted a VPN where the main action is one tap and the connection is fast enough that you don't feel tempted to turn it off.

How V Shield helps

V Shield creates an encrypted connection between your device and secure VPN servers, which helps protect your data on public and private networks. You connect or disconnect with a single tap and can choose from high-performance servers around the world for browsing and streaming.

It supports the modern WireGuard and Shadowsocks protocols for fast, stable connections.

Key features

One-tap connect and disconnect
Global server locations
WireGuard and Shadowsocks protocol support
Privacy-focused design that doesn't monitor, record or store your browsing activity
Connection statistics: server location, session duration and data usage
iOS Live Activity so you can follow your connection status

An honest note

A VPN is useful, but it isn't magic. It encrypts traffic between your device and the VPN server, which matters most on untrusted networks. It doesn't make you anonymous, and it doesn't replace good security habits like strong passwords and keeping software updated. I'd rather say that plainly than overpromise.

Also, some features require a Pro subscription. You can find the details in the app.

Who it's for

People who regularly use public Wi-Fi
Travellers who want to protect their traffic on the go
Anyone who wants an extra layer of online privacy without a complicated setup
Users who care which protocols their VPN uses

Try it

Feedback welcome

Performance varies a lot by region and network, so real-world reports are really valuable:

How is connection speed and stability where you are?
Which server locations would you like added?
Is the Live Activity on iOS useful to you?

If you've worked with WireGuard or Shadowsocks on mobile, I'd like to hear your experience too. Leave a comment below!

I Built StoryFlow: From a Text Prompt to a Finished AI Story Video

Slayer King — Sun, 25 Oct 2026 12:55:11 +0000

I'm the developer of StoryFlow Auto Story Generator, published by my company CYBORGATE LTD (https://cyborgate.uk). Here's what it does and why I made it.

The problem

Making a short story video usually takes several skills. You need to write a script, design characters, find or create visuals, then edit everything with pacing and transitions. Lots of people have good story ideas but don't have the time, tools or editing experience to turn them into videos. I wanted to see how far a single prompt could get you.

How StoryFlow helps

You describe your story, for example "a detective cat in Paris" or a sci-fi thriller. StoryFlow drafts a structured, scene-by-scene screenplay with dialogue, narration and shot directions.

You stay in control. You can review the script, edit it in the script editor, or regenerate it with one tap. Then StoryFlow generates character artwork in a consistent style and combines the script, characters and cinematic visuals into a finished video with scenes, transitions and pacing. You can download or share it.

Key features

Prompt to screenplay: scene-by-scene scripts with dialogue, narration and shot directions
Script editor and one-tap regeneration
Consistent character artwork generated for your story
Finished video with scenes, transitions and pacing
Durations: 1, 3 or 5 minutes
Visual styles: Normal, Anime, Cinematic, Fantasy, Horror and Noir
Aspect ratios: 16:9, 9:16 or 1:1 for YouTube, TikTok or Instagram
Discover feed to browse what others create
No editing skills, templates or stock footage needed

An honest note

StoryFlow is iOS only for now. There's no Android version yet. Like any generative AI tool, the output depends on your prompt, and some results will be better than others. That's why the script review and regenerate step exists. I'd treat it as a fast way to prototype and produce story content, not as a replacement for a professional animation studio.

Who it's for

Content creators who need story videos for YouTube, TikTok or Instagram
Writers who want to see their ideas visualised
Hobbyists who enjoy storytelling but don't edit video
Anyone curious about what AI story generation can do today

Try it

App Store: https://apps.apple.com/gb/app/storyflow-auto-story-generator/id6782660912
Google Play: not available yet

Feedback wanted

This is the app where I'm learning the most, and your input would really help:

Which visual style gave you the best results?
How well does character consistency hold up across scenes?
Would an Android version matter to you?
Which durations or formats are missing?

If you're building with generative video or image pipelines, I'd like to hear how you handle consistency and pacing. Comments are open!

I Built UmmahPlus: Prayer Times, Qibla, Quran and a Ramadan Tracker in One App

Slayer King — Sun, 25 Oct 2026 12:53:30 +0000

I'm the developer of UmmahPlus Namaz Ramdan Tracker, published by my company CYBORGATE LTD (https://cyborgate.uk). Here's why I built it and what it includes.

The problem

Many Muslims use one app for prayer times, another for the Qibla, another for reading the Quran, a separate tasbih counter, and then something else again during Ramadan to track fasts. Many of these apps are cluttered or full of ads. I wanted one calm, well-designed companion that supports daily practice without getting in the way.

How UmmahPlus helps

UmmahPlus calculates accurate prayer times for your location and sends customisable Adhan notifications. A built-in compass shows the Qibla direction wherever you are. Each day starts with curated Quran verses, Hadiths and supplications. You can read the Quran in the app and keep count of dhikr with a digital tasbih.

During Ramadan, you can log your daily fasts, track Taraweeh prayers and keep on top of your goals.

Key features

Accurate, location-based prayer times
Customisable Adhan notifications
Qibla compass
Ramadan tracker for daily fasts, Taraweeh and goals
Daily verses, Hadiths and supplications
In-app Quran reader
Digital tasbih

UmmahPlus Pro

Pro is available as a monthly or yearly subscription. The yearly plan includes a 3-day free trial. Pro unlocks:

An unlimited AI Islamic assistant for questions, verse context and Islamic history
Premium Quran recitations from renowned Qaris
A completely ad-free experience
Advanced statistics on your Quran reading streaks, prayer habits and fasting history

An honest note

The AI assistant is there to help you explore context and history. It doesn't replace qualified scholars. For rulings and personal religious questions, please consult knowledgeable people you trust. I also want to be transparent that the free version shows ads and that removing them is part of Pro.

Who it's for

Muslims who want reliable prayer times and Adhan reminders
Travellers who need the Qibla direction in unfamiliar places
Anyone who wants to build a daily Quran reading habit
People who want to track fasting and Taraweeh during Ramadan

Try it

App Store: https://apps.apple.com/gb/app/ummahplus-namaz-ramdan-tracker/id6759190731
Google Play: https://play.google.com/store/apps/details?id=com.cyborgate.ummahplus

I'd love your feedback

Prayer time accuracy matters a lot, and I'd like to know how it performs for you:

Are the prayer times accurate for your location and calculation method?
Is the Qibla compass reliable on your device?
What would make the Ramadan tracker more useful?

Developers, if you've worked with location-based calculations or compass sensors, I'd be glad to compare notes. Please leave a comment!

I Built Easy Expense: A Receipt Scanner for Freelancers and Small Businesses

Slayer King — Sun, 25 Oct 2026 12:52:01 +0000

I'm the developer of Receipt Scanner – Easy Expense, published by my company CYBORGATE LTD (https://cyborgate.uk). This post covers what it does and who I built it for.

The problem

At tax time, many freelancers and small business owners end up with a shoebox, a drawer or a glovebox full of faded paper receipts. Some are lost, and the ones that are left still need to be typed into a spreadsheet. I wanted a tool where you capture a receipt once, at the moment you get it, and never have to deal with it again.

How Easy Expense helps

Scan a paper receipt with your camera or upload a photo. OCR detects the edges, crops the image and extracts the merchant, date, amount and more. You can edit any field later, because OCR isn't perfect and you should always have the final say.

From there, you organise expenses by category and vendor, group them into reports for projects or clients, and export the data when your accountant needs it.

Key features

Camera scanning or photo upload
OCR with automatic edge detection and cropping, extracting merchant, date, amount and more
Editable fields so you can fix anything
Business and personal expenses, organised by category and vendor
Expense reports for projects or clients, with simple spending summaries
CSV export to share with your accountant or team
Secure cloud backup and sync across devices

Premium

Premium adds:

Unlimited scanning
Email receipt importing
Bank and credit card statement scanning
Multiple businesses
Team collaboration
Priority support

An honest note

Easy Expense is an organisation tool, not an accounting or tax service. It helps you capture, categorise and export your records so you or your accountant can file properly. I deliberately kept it simple. If you need full bookkeeping, use it alongside your accounting software rather than instead of it.

Who it's for

Freelancers who need to keep track of deductible expenses
Contractors and consultants who bill expenses back to clients
Small businesses that want receipts in one place, all year round
Anyone who has ever spent a weekend rebuilding a year of expenses from memory

Try it

Feedback, please

I'd really appreciate hearing:

How well does OCR handle your receipts (crumpled, faded, foreign currency)?
Does the CSV export fit what your accountant expects?
What's missing from your expense workflow?

If you've worked on OCR or document scanning, I'd also like to hear what worked for you. Drop a comment below!

I Built Pregsi: One App for Trying to Conceive, Pregnancy and Period Tracking

Slayer King — Sun, 25 Oct 2026 12:35:00 +0000

I'm the developer behind Pregsi, an app published by my company, CYBORGATE LTD (https://cyborgate.uk). Here's what it does, who it's meant for, and where I'd like your feedback.

The problem

Many people use several apps to cover what is really one journey. There's a period tracker, then an ovulation calculator when they start trying to conceive, then a pregnancy app with weekly updates, and later a kick counter and a contraction timer. Each switch means starting your data over and learning a new interface. I wanted one companion that covers trying to conceive, pregnancy and late pregnancy in a single place.

How Pregsi helps

Pregsi is built around three stages.

Trying to conceive. Pregsi estimates ovulation dates and your fertile window on an ovulation calendar. You can log basal body temperature and cervical mucus, and set fertility reminders.

Cycle tracking. The period tracker records period dates, flow, cycle length, moods and symptoms, and reminds you when things are due.

Pregnancy. You can follow your pregnancy week by week, with fetal growth updates, baby size comparisons and milestones. The part I enjoyed building most is the interactive 3D view of your developing baby.

Key features

Week-by-week pregnancy tracking with growth updates, size comparisons and milestones
Interactive 3D baby views
Logs for symptoms, weight and hydration
Prenatal appointment organiser, ultrasound notes and medication reminders
Ovulation calendar and fertile window estimates
Basal body temperature and cervical mucus logging
Period and cycle tracker with mood and symptom logs
Kick counter and contraction timer (records duration and frequency) for later in pregnancy
Reports that summarise your logs over time

An honest note

Ovulation predictions are estimates. They are not a form of contraception. Pregsi is meant to help you organise information and notice patterns. It doesn't replace medical advice, so please talk to a healthcare professional if you have any concerns. I put this caveat inside the app too, because I think a health app should say clearly what it can and can't do.

Who it's for

People tracking their cycles who want clear records and reminders
Couples trying to conceive who want fertile window estimates and fertility logging in one place
Expecting parents who want weekly updates, appointment organisation and a 3D look at their baby
Anyone in late pregnancy who wants a simple kick counter and contraction timer

Try it

App Store: https://apps.apple.com/gb/app/pregsi-pregnancy-ovulation/id6763647859
Google Play: https://play.google.com/store/apps/details?id=com.cyborgate.pregsitracker

I'd love your feedback

I'm still improving Pregsi, and real feedback helps more than anything else. If you try it, please tell me:

What felt confusing or slow?
Which feature did you expect that isn't there?
Does the move from cycle tracking to pregnancy mode feel smooth?

Leave a comment below. Developers who've built health or tracking apps, I'd also like to hear how you handle reminders and data presentation. Thanks for reading!

I Built ULO AI: An AI Tutor for Actually Speaking English

Slayer King — Sun, 25 Oct 2026 12:34:20 +0000

I'm the developer of Speak & Learn English: ULO AI, published by my company CYBORGATE LTD (https://cyborgate.uk). Here's the problem it tries to solve and how it works.

The problem

Lots of English learners can read and write reasonably well but freeze when they have to speak. Grammar drills and vocabulary flashcards help, but they don't give you practice in real conversation. A human tutor does, but tutors cost money and aren't available at 11pm before a job interview. I wanted learners to have a patient conversation partner they can use whenever they need one.

How ULO AI helps

ULO AI is a personal AI English tutor built around conversation. You can speak naturally or type your replies, and you'll hear natural AI voices in response. As you talk, you get pronunciation feedback and gentle grammar corrections. The aim is to help you keep the conversation going, so corrections are kept gentle.

Key features

Speak or type: voice or text conversation, with natural AI voices
Pronunciation feedback and gentle grammar corrections during the conversation
Adaptive lessons from beginner to advanced
Daily learning roadmap, progress tracking and learning streaks
Vocabulary building and grammar guidance
Role-play scenarios: job interviews, business meetings, airports and travel, hotels and restaurants, shopping, doctor appointments, school, phone calls, customer service and social situations
Instant translation between supported languages, with AI explanations of grammar, vocabulary and natural expressions
Ask the tutor anything about a phrase, rule or pronunciation

Why role-play matters

I think the role-play scenarios are the most useful part. If you have a job interview next week, you can practise the interview itself. If you're travelling, you can practise checking into a hotel or ordering in a restaurant. Going through a situation once before it happens makes a real difference to your confidence.

An honest note

An AI tutor is a practice tool. It doesn't replace immersion, real conversations with people, or a qualified teacher for exam preparation. What it gives you is unlimited, low-pressure practice that's always available, and for many learners that's the part they're missing.

Who it's for

Learners preparing for work, study, travel or exams
People who understand English but lack confidence when speaking
Professionals rehearsing interviews, meetings or phone calls
Anyone who wants daily, structured practice with streaks to keep them going

Try it

App Store: https://apps.apple.com/gb/app/speak-learn-english-ulo-ai/id6790012817
Google Play: https://play.google.com/store/apps/details?id=com.cyborgate.aitutor

Tell me what you think

I'm actively improving ULO AI and would value your input:

Which role-play scenario should I add next?
Does the pronunciation feedback feel useful or too picky?
Is the level adaptation right for you?

If you're building in language learning or speech tech, I'd like to hear your experience too. Comments are open!

I Built CAL AI: Calorie Tracking Without the Tedious Manual Entry

Slayer King — Sun, 25 Oct 2026 12:33:27 +0000

I'm the developer of CAL AI – Calories Tracker, published by my company CYBORGATE LTD (https://cyborgate.uk). In this post I explain why I built it, what it does, and what it doesn't do.

The problem

Calorie counting works for many people. The hard part is logging. You search a database, choose the right item, guess the portion and repeat that for every meal. Most people I know gave up on tracking because typing it all in got tedious, not because they lost interest in their goals. I wanted to make logging fast enough that people would keep doing it.

How CAL AI helps

CAL AI tracks calories, protein, carbohydrates and fats through the day. It shows your progress against personalised daily goals, which are based on your body information, activity level and objective. That objective might be losing weight, maintaining, building muscle or just understanding your eating habits.

The time saver is in CAL AI Pro. You take a photo of a meal, and AI identifies the foods and estimates their nutritional content. You check the result and log it, without typing anything.

Key features

Tracking for calories, protein, carbs and fats
Personalised daily goals based on your body, activity and objectives
Weight goals and daily nutrition progress on a clear dashboard
Pro: AI photo food scanning
Pro: unlimited meal logging
Pro: AI-generated nutrition insights and advanced progress features

Pricing

The free version gives you a limited logging experience. Pro is an in-app purchase with monthly, yearly or lifetime options. I'd rather say that upfront than have you find out after installing. The core dashboard and goal tracking are free to try, and the AI scanning and unlimited logging are paid.

An honest note on AI estimates

Photo scanning produces estimates. AI can identify foods and approximate their nutrition, but portion sizes and hidden ingredients like oils and sauces are hard to judge from a picture. Treat the numbers as a fast, reasonable starting point, and adjust them when you know better. If you have specific medical or dietary needs, check with a professional.

Who it's for

People who want to lose weight, maintain or build muscle and need to see their calories and macros
Anyone who has tried calorie counting before and stopped because logging was tedious
Gym-goers who mainly care about hitting a protein target
Curious people who want to understand their eating habits without obsessing over every entry

Try it

App Store: https://apps.apple.com/gb/app/cal-ai-calories-tracker/id6800268062
Google Play: https://play.google.com/store/apps/details?id=com.cyborgate.calai.caltracker

Feedback welcome

I'd really like to hear from people who use it:

How accurate did the photo scans feel for your typical meals?
Is the dashboard clear at a glance?
What would make you log more consistently?

If you're a developer working on vision or food recognition, I'd also like to swap notes. Leave a comment and I'll read every one.

How I Built and Shipped 9 Mobile Apps (and What I Learned)

Slayer King — Sun, 25 Oct 2026 12:30:23 +0000

Disclosure: I'm the developer behind all of these apps at CYBORGATE LTD. This is a #showdev post, so feedback, including the harsh kind, is welcome.

A while ago I made a simple rule for myself: ship small, focused apps and let real users tell me what's worth doubling down on. Many months later, there are nine apps live across the App Store and Google Play. This post covers what they are, how I approached them, and what I'd tell anyone thinking of doing the same.

The portfolio

App	What it does	Links
Pregsi	Pregnancy week by week with 3D baby views, ovulation and period tracking	iOS · Android
CAL AI	Photo-based calorie and macro tracking	iOS · Android
ULO AI	AI English speaking tutor with role plays	iOS · Android
Easy Expense	Receipt OCR, expense reports and CSV export	iOS · Android
UmmahPlus	Prayer times, Qibla, Quran, Ramadan tracker	iOS · Android
StoryFlow	Text prompt → script → characters → finished story video	iOS
V Shield VPN	One-tap VPN with WireGuard and Shadowsocks	iOS · Android
VPN MAX Pro	Simple, privacy-focused one-tap VPN	iOS · Android
DeepSearch	AI search across public info on people, companies and topics	Android

1. Pick problems people already search for

Every app here targets a need people already type into a store search bar: "calorie counter", "receipt scanner", "prayer times", "VPN". I didn't try to invent categories. I tried to make a cleaner, more focused version of things people already want. Discovery is the hardest part of mobile, and building into existing demand makes it a lot easier.

2. One core loop per app

Each app has one thing it must do well:

CAL AI: photo → nutrition estimate → progress toward goal
Easy Expense: scan → extract merchant/date/amount → export
ULO AI: speak → feedback → try again

Everything else is secondary. Whenever I was tempted to add a feature, I asked whether it made the core loop faster or clearer. If it didn't, it went into the backlog.

3. AI where it removes friction, not as decoration

Several of the apps use AI: food recognition in CAL AI, conversation and corrections in ULO AI, script and character generation in StoryFlow, the Ask AI feature in DeepSearch, and an assistant in UmmahPlus Pro. The test I used each time: does AI remove a tedious step the user hates? Typing every ingredient into a food log is tedious, and so is staring at a blank page when you want to make a story video. Those are the places AI earns its spot.

I'm also careful about how I describe what it does. Nutrition values are estimates, ovulation predictions in Pregsi are estimates and not contraception, and DeepSearch surfaces only publicly available information. Honest copy builds trust, and app reviewers care about it too.

4. Monetisation: be upfront

Most of the apps are free to start, with a Pro tier for heavier use (unlimited scans, ad-free use, premium content). I've learned to state clearly what's free and what isn't. Users who feel tricked leave one-star reviews, and they're right to.

5. The store listing is half the product

For a small developer, screenshots, the first two lines of the description and the icon do more marketing than anything else. I rewrite listings far more often than I expected. Clear benefit-first copy ("Snap a meal, track calories and macros") beats a feature list every time.

6. Shipping on both platforms

Seven of the nine apps are on both iOS and Android. StoryFlow is iOS-only and DeepSearch is Android-only for now. Launching on one platform first, learning, and then expanding has worked for me when time is tight.

What's next

I'm putting most of my effort into the apps users actually come back to, based on reviews and retention, not on which ones I personally like most.

I'd love your feedback. Pick any app, try it for five minutes, and tell me in the comments what confused you or what you'd change. Everything lives at cyborgate.uk.

I built a Sovereign, Self-Hosted Personal AI Agent Workspace: Meet Poka 🤖🛡️

Slayer King — Thu, 01 Oct 2026 14:46:10 +0000

Why I built Poka — an open-source, local-first alternative to cloud AI agents with governed tool approvals, multi-model support, and a sandboxed browser runtime.
tags:

As autonomous AI agents continue to evolve, most of the industry is heading in one clear direction: closed cloud platforms.
Services like OpenAI Operator, Claude Cowork, Manus, and Meta Muse promise seamless agentic workflows. But they also require you to surrender:

Your Data Sovereignty: Conversation history, sensitive files, and execution logs live on remote corporate servers.
Your API Keys & Secrets: Credential management is completely opaque.
Execution Safety: Agents often execute arbitrary shell actions or web tasks in environments you cannot inspect or audit. I wanted something different: a sovereign, inspectable, self-hosted personal AI agent workspace where you retain absolute ownership over your data, credentials, and execution environment. That is why I created Poka.

🌟 What is Poka?

Poka is an open-source, local-first personal AI agent platform built with Next.js 15 (React 19), FastAPI, SQLite, and Docker/Playwright.
It combines multi-model conversations, an action gateway with human-in-the-loop approvals, external app connectors, and an optional sandboxed desktop runtime into one unified workspace.

💡 The Core Pillars

1. 🛡️ Governed Action Gateway (Deny-by-Default)

The biggest hazard of autonomous agents is letting them run unvetted terminal commands, script executions, or file modifications.
Poka implements a governed action gateway:

Low-risk operations (workspace reads, web search) execute automatically and are recorded in a tamper-resistant audit trail.
High-risk operations (shell execution, writing files, sending OS keystrokes) automatically pause execution and prompt for explicit user approval. You can inspect the exact payload, command, or script before the agent is permitted to run it. ### 2. 🔒 Absolute Privacy & Local-First Storage
All conversations, assistant personas, artifacts, and logs reside in a local SQLite database (~/.poka).
Provider API keys and secret tokens are encrypted with 256-bit Fernet keys at rest.
Browser authentication uses server-enforced HttpOnly session cookies — no master tokens are ever exposed to the frontend JavaScript runtime. ### 3. 🤖 Universal Multi-Model Support (Cloud & Local) Poka isn't tied to any single AI provider. You can switch between models on the fly:
Cloud Providers: OpenAI, Anthropic Claude, DeepSeek, Groq, Mistral.
Local Inference Engines: Connect directly to Ollama, vLLM, or LM Studio via OpenAI-compatible endpoints with zero external network leakage.
Multiple protocols supported: standard /chat/completions, streaming /responses, and direct prediction endpoints. ### 4. 🖥️ Sandboxed Computer Automation Need the agent to navigate the web or automate desktop workflows?
Docker / Playwright Sandbox: Runs an assistant-scoped container with a 30 FPS display stream, virtual input, read-only root filesystems, and dropped Linux capabilities.
Windows Desktop Agent: Includes a dedicated agent script for remote PowerShell execution and native Windows GUI automation with DPAPI token security. ### 5. 🔌 Ecosystem Connectors & Keyless Search
Keyless Web Search: Real-time web browsing via a built-in You.com MCP adapter (/search <query>) that works out-of-the-box without requiring an API key.

- Composio App Connectors: Plug into GitHub, Slack, Notion, Linear, Google Docs, and Sheets with bounded OAuth scopes.

🚀 Stop Trusting "Top 10 AI Tools" Lists. Find AI Tools Using Real Public Reviews Instead.

Slayer King — Sun, 19 Jul 2026 17:59:14 +0000

Every day, a new AI tool launches.

A new chatbot. A new coding assistant. A new AI agent. A new image generator.
Within weeks, there are hundreds of blog posts claiming "This is the #1 AI Tool."

But here's the problem...
❌ Many rankings are sponsored.
❌ Most reviews simply repeat marketing copy.
❌ Comparing AI tools means opening 20+ browser tabs.
Finding the right AI tool shouldn't be harder than using one.
That's why I built AgentHiveX.

🔍 What is AgentHiveX?
AgentHiveX is an independent AI Tools Directory that helps developers, founders, creators, and businesses discover the best AI products using real public reviews, community ratings, and open-source signals—not marketing hype.

Whether you're searching for:
🤖 AI Agents
💻 Coding Assistants
🎙️ Voice AI Platforms
🎨 Image & Video AI
⚡ Automation Tools
📚 Open-Source AI Projects
🛠️ Developer APIs
📈 Productivity Tools
...AgentHiveX helps you compare them in one place.

🌐 Explore AgentHiveX
If you're tired of sponsored rankings, clickbait articles, and spending hours comparing AI products...

👉 Check out AgentHiveX: https://agenthivex.com