The Free Tools
Playbook

The job: you need a hero illustration, a mood board image, or thirty concept variants — today, without a subscription. The play: open LND AI's Text to Image and type. The tool generates from a text prompt with 12 art styles, 9 image sizes, and 1–4 images per batch, plus a seed field so a good result can be reproduced later instead of being a one-off lottery ticket. Full-resolution JPG download, no watermark.

1. Write the prompt as a description, not a wish: "a rainy chai stall in Mumbai, warm lantern light, cinematic" beats "nice picture of tea".
2. Pick a style that matches the medium (photo vs illustration vs anime) and the aspect ratio before generating — batches inherit it.
3. When a result is 90% right, reuse its seed and tweak one phrase at a time to converge.

Why this over the big names: Midjourney wants a subscription and a Discord account; this is a URL. The trade is honest and stated in LND's long-form guide on how diffusion models work: free daily credits mean a queue and a daily budget, not infinite generations. Credits refresh automatically at midnight IST, every day, with no account needed — an optional free account (username and password, nothing else) bumps the refresh and adds a one-time signup bonus. The guide even teaches prompt-writing and honest limitations, which is more hand-holding than most paid tools give.

The job: a product photo with a kitchen in the background, a logo stuck at 480px, a square shot that needs to become a banner. Four separate jobs, one tab: the Background Remover does pixel-precise cutouts (hair and fur included — the honest hard case) and hands you a transparent PNG; the Background Replacer swaps in a whole new scene from a text prompt, with automatic relighting and colour-match on the subject; the 2× Image Upscaler sharpens low-res images for print; and Uncrop / Extender generates matching scenery around the edges so you can change aspect ratios without the dreaded stretch.

1. Chain the tools in order: cut out → replace background → upscale. Each output downloads clean and feeds the next.
2. For social-media resize jobs, skip Uncrop and jump to Onyx's preset resizer instead (Play 3) — fixed ratios are one click there.
3. Keep the original file; browser tools are fast to iterate, so re-runs cost nothing.

The job: crop, filter, blur a face, add text with a real font, mask into a circle, convert HEIC to JPG — the everyday 80% of Photoshop, in a browser, forever free, no signup, no watermark. Onyx is a privacy-first editor where all processing runs locally — your photos never leave your device, which makes it the only editor in this playbook I'd hand a folder of family pictures.

The tool deck: crop (with platform presets), resize, filters, blur, grayscale, flip, add-text with 40+ Google Fonts, masking (circles, hearts, stars), shapes, freehand draw, rotate, and format conversion across PNG, JPG, WebP, SVG, BMP, GIF and HEIC. Social teams should bookmark the resize presets — Instagram story, post, YouTube thumbnail, Twitter header, Facebook cover, LinkedIn banner, passport photo — and the dedicated converters like HEIC→JPG for every iPhone owner on Earth.

Onyx publishes three deep comparisons worth reading before you commit anywhere: Onyx vs Canva, vs Photopea and vs Pixlr, plus a tutorial blog at onyx.namansoni.in/blog (cropping guides, JPG-vs-PNG-vs-WebP, 2026 social image sizes).

The job: a scroll-stopping 3D background clip, or a birthday flex where someone's name falls out of the sky as physical objects. 3Dfy generates infinite AI 3D worlds you can fly through and render into film, with camera direction for frame-perfect shots. The full studio at /3dfy builds anything from scratch; /make3d turns prompts into importable AI shapes and worlds; and the name playground at /world turns your name into living 3D physics — letters you can knock over, stack and re-arrange.

1. Start in the world generator to learn the vocabulary of scenes that work ("neon canyon at dusk", "candy hills").
2. Move to the studio when you need the camera path exactly right — direct, don't hope.
3. Render short loops; 6–10 second clips loop cleanly on social.

The job: an instrumental for a cover, a clean acapella for a remix, practice tracks for the band. Monyx Studio runs an AI vocal remover on MDX-Net models entirely in your browser — the audio never leaves your device, and there's no signup, no watermark and no per-track fee. Upload a song, get separated vocals and instrumental, download both.

Moises — the app most musicians know — is genuinely good, but it meters free separations and wants a subscription for the full toolkit. Monyx's bet is the opposite: your laptop's CPU is the meter, so the tool is free and private, and separation quality on heavily-produced pop is strong (live recordings and dense wall-of-sound mixes remain the honest hard cases for every separator, including the paid ones).

The job: actually sing — and record it without sounding like you're in a bathroom. Monyx's Karaoke Studio is the killer feature: upload a song, trim it, isolate the backing music with the AI from Play 5, then sing along with real-time acoustic echo cancellation (the room stops fighting you), and merge your clean vocal with the BGM as a finished track. For everything else there's a full multitrack DAW with timeline, FX rack and mixer; virtual instrument pads — piano, guitar, drums, bass, pads — playable and recordable live; and a video maker that stitches images or clips with your audio into an MP4 with text overlays and fades.

1. Karaoke first: song → trim → AI isolate → sing with echo cancellation on → merge. Headphones, always.
2. Then pads for a quick original bed, and the DAW to arrange: your vocal on track 1, BGM on 2, pad loop on 3.
3. Finish in the video maker if the song needs a shareable face.

The job: a birthday/anniversary/friendship gift that isn't a mug. Gaana Banana is India's personalised AI song maker: write a name, pick a theme — romantic, birthday, yaari, family, wedding, festival, motivation, funny, bhakti, sorry/thanks — and in 1–2 minutes you get a full song with vocals, MP3 download and WhatsApp-ready share. Bring your own lyrics (up to 800 characters) or let the AI write them.

This is the one tool in the playbook that is paid per output, and the pricing is refreshingly flat: ₹28 = 1 song credit, ₹45 = 2, ₹175 = 10, paid over UPI. Credits live on a 12-digit key you can transfer between devices, and — the detail that earns the honesty score — failed generations refund automatically. The full walkthrough (including the AI se gaana kaise banaye guide, in Hinglish) is on the site, with plans at /pricing and answers at /faqs.

The job: a Hindi voiceover for a reel, a lecture recording that needs to become text, a paragraph translated to Tamil. Three LND tools, one tab. Text to Speech covers 22 Indian languages with 40+ voices and 0.5×–2× speed control — and its built-in Browser engine is always completely free. Speech to Text transcribes files or live microphone audio with two engines: a high-accuracy server engine, and a cheaper browser engine that runs offline on your device. Translate handles English ↔ 22 Indian languages up to 10,000 characters a shot.

1. Script in short sentences; TTS pauses where you'd pause.
2. Pick the Browser TTS engine for drafts (free, instant) and the HQ engine for the final render.
3. For lectures: record on your phone, upload the file, transcribe, then paste key paragraphs into Translate.

The job: your track deserves a video and you have zero video skills. Drop any MP3 on Dance Cat 3D and a fully procedural 3D cat dances in perfect beat sync. The engineering is the fun part: real-time Web Audio FFT analysis, offline BPM detection, and spectral-flux kick analysis — the cat genuinely hears the kicks — and it all runs 100% in your browser: no AI, no uploads, no server. Your file never leaves the machine.

Screen-record the best 15 seconds and you have a loop for a reel. Kids and cats-in-general audiences are guaranteed.

The job: 2AM, and you want to talk — not to a search box with a personality skin, but to a character who knows your name and remembers what you told them last week. Talk is a free AI-companion app built for India: 11 companions, no sign-up, no OTP — type a name and you're in. Hinglish-first with a pure English mode, installable as a PWA.

The difference from a chatbot is the 4-tier memory system (running summaries, hard facts, an event log, and semantic recall) plus live voice calls with Indian voices, barge-in and call memory — they pick up the phone, every time. The cast covers the emotional spectrum: Arya the 2AM bestie, Kabir the startup mentor, Mira and Aarohi for girlfriend energy, Rohan the hype man, Akansha's deadpan comedy, Rhea's chaos, Rupali the literature-teacher crush, Ruhi the desi padosan from flat 402, Noor the midnight RJ — and Gojo Satoru for anime fans (fan-inspired, unofficial, unaffiliated with Shueisha/Gege Akutami/MAPPA). Meet everyone at /companions.

Free accounts get 300 credits daily (~50 messages), refreshed automatically; optional packs start at ₹29 and unlimited plans skip metering. An 18+ Spicy Mode exists — one-time unlock from ₹9, off by default, consent-first — and the platform is unusually straight about it in its Spicy Mode explainer and its honest guide on whether AI companionship is even good for you. That's the honesty score, right there.

The job: a commute that isn't doom-scrolling. Talk's five interactive stories are chat-fiction you star in: cinematic narration, characters who talk to you and each other, choices that move trust meters and change scenes. Your decisions decide the ending — and every world has an endless mode after the finale.

• My First Day in College — a 50-day college romance where you're the new kid.
• Kokuhaku Club — anime rom-com: save a dying confession club in 14 days.
• Last Show, First Row — save a single-screen cinema in 14 days (all-adult cast).
• The Anniversary Heist — family heist comedy: bhai, a vase, and damage control.
• Love All — 28 days of 5AM badminton and a mixed-doubles partnership with Ira Deshpande.

Each ends in acts with real structure — this is written drama with a director engine, not a random chat wrapper. If interactive fiction is your genre, start with Love All; if anime, Kokuhaku Club.

The job: ten quiet minutes against a bot that doesn't spam you with premium. Chess — Human vs Bot is a clean, fast browser board: pick your colour, pick the bot's thinking time (Balanced sits at ~1.5 seconds a move), play. Promotion is handled properly, and the move history lets you replay what the bot did to you — which is how you actually improve.

No rating pressure, no ads, no "unlock the analysis engine". It's a sparring partner, not a casino. Pair it with the free chess openings guide for club players from this same network when you want the studying half.

The job: the idea in your head, live on the internet, tonight. NewEra is a free online AI agent — no login, no card — that takes a plain-English description, writes the files, pushes a GitHub repo, runs a real CI build on GitHub Actions (4 vCPU / 16 GB runners), reads its own failure logs, fixes them, and deploys the green build to Cloudflare Pages (EdgeOne as fallback). You watch every stage and end with a live URL. This is not a sandbox preview — it's the actual pipeline.

The specs read like a startup's infra deck: up to 4 parallel sub-agents, up to 12 hours of VM build time with auto-relay, live web search and current library docs for research, in-browser runtimes (Python via Pyodide, SQLite via WASM, FFmpeg, Git), and every project exports as a single portable .newera file. Privacy is handled like an adult: repos are created only with explicit consent, and self-clean repos auto-delete. The workspace lives at /agent.

1. Describe the app like you'd brief a freelancer: pages, data, one clear job.
2. Watch the build; if CI fails, let the agent read the logs before you panic.
3. Export .newera when it works — your project, your file, portable forever.

The job: a coding agent you control — including the bill. Pi in the Browser runs the real Pi coding agent fully in your browser via WebAssembly, spending your own Pollen on Pollinations models. Bring your account, pick a model, and let it build — the sandbox is browser-hosted, so the compute story is unusually clean: your key, your meter, no middleman markup.

Think of it as the BYO version of Play 13: NewEra is the all-inclusive agent, Pi is the agent for people who already have model accounts and want direct control. The free-tier score reflects that BYO means it's free software, not free compute — the honest framing is on the site's own description, which I appreciate.

The job: six accounts, one phone, and the creeping feeling that SMS codes aren't enough. 2FA Authenticator is a free, browser-based two-factor app: paste your TOTP/HOTP setup keys, get time-based codes on desktop or mobile, and choose your storage model — keep keys on your device only, or sync to an encrypted account when you want a new-machine recovery path. The classic Google Authenticator trap (lose phone, lose everything) is a conscious design decision here, not an accident you discover later.

The desktop angle is underrated: most authenticators are phone-first, which means typing 6 digits you can half-see. This one lives where you're already typing. For the deep version, read the network's free TOTP 2FA guide — same ecosystem, full how-to.

The job: every shop, society and mohalla runs on udhaar — and the tracking runs on memory and arguments. Khata is the ecosystem's Hinglish-first digital ledger: a PWA you add to your home screen, with chat and voice entry — say or type "Sharma uncle ko 250", the way you'd actually say it. The desi answer to the notebook behind the counter, from the same maker as everything else here. (Full disclosure: it was mid-redeploy at compile time — if the link blinks, it's infrastructure mood, not abandonment; the sibling digital khata guide from this network covers the method in depth.)

The job: a bio link that isn't an ad farm. Links gives every person one fast page — your links, mini descriptions, contacts and photo — with no ads, no trackers, no analytics, on free infrastructure. The creator's own page at links.namansoni.in/namansoni78 is the live demo.

The feature nobody else has: embeddable SVG widgets. Build a self-contained card in the widget builder at /widgets — with an AI-assisted interview flow that turns your answers into a finished widget — then embed it in a GitHub README or any Markdown page via ![](https://links.namansoni.in/api/embed/<slug>). The embed API is civil engineering: width/height parameters (120–4096px) never distort the card, a v cache-buster mints on every save so READMEs refresh within minutes, and responses are CDN-cached for a day.

The job: the small stuff that eats your day — a QR code, a strong password, a file converted, a Wi-Fi speed check. The More Tools hub is 20+ browser utilities, and here's the line that makes it remarkable: utilities never use credits and are always completely free. Not "free tier" — free, full stop.

The drawer: QR generator, password generator, hash converter (SHA-256/512, MD5, Bcrypt, PBKDF2), binary/hex converter, colour and unit converters, text encode/decode, scientific calculator, world clock, sound meter, GPS speedometer, spirit level, inclinometer, IP detector, net speed test, QR/barcode scanner, live HTML viewer, notes app, online whiteboard, audio recorder, weather check, image resize/convert/ASCII/SVG/OCR, tone generator, coin toss, CPS test.

Phone as a toolbox is the sleeper hit: spirit level for the shelf, speedometer on the dash, sound meter for the neighbour conversation, dB meter for the generator. Each utility has its own chapter in the Learn library — 43 long-form guides in total.