Privacy choices

Optional Google Analytics and advertising are off until you choose. Read our privacy details.

ISSUE NO. 107AI + DEV TOOLS
AutoKaam.

AutoKaam, AI tools, field-tested by an Indian operator

Issue
№ 107
Stories
101
Latest
08/05
Scoops
9
$ in Play
$1784.3B
Status
VERIFIED
FIELD NOTEVERDICT · AUG 5, 2026 · 7 MIN

303 Routes, 0 Red Gates: Static Export Still Needs a Preflight

A green build is not release clearance. The artefact needs its own route, metadata, link, and visual checks.

AutoKaam Editorial··NEXT.JS / STATIC EXPORT
COVER · BUILD OUTPUT + CRAWLERS

The build output is the deployment candidate.

SOURCE: AutoKaam release preflight

303
opening baseline routes
BUILD OUTPUT + CRAWLERS
Named stake
What AutoKaam Thinks
  • Treat out/ as the release candidate.
  • Fail the export when an indexable page loses its H1 or social image.
  • Check a dynamic route from the built tree, not only next dev.
  • Publish the evidence with the hand-off, not after a crawler finds the miss.

From the Playbook

Operator-tested tools, with INR pricing.

View all 30

Claude.ai Pro: ~Rs 1,700/mo (USD 20). Claude Max (5h sessions): ~Rs 8,500/mo (USD 100). API Sonnet 4.6: USD 3 input / USD 15 output per 1M tokens. API Opus 4.7: USD 15 input / USD 75 output per 1M tokens.

Claude, Anthropic's Sonnet and Opus Families

The model I reach for when reasoning matters more than throughput.

Free tier available with Claude Pro. Claude Max (recommended for empire-scale work): ~Rs 8,500/mo. Pay-as-you-go API also supported.

Claude Code, the CLI Agent I Run All Day

Anthropic's terminal-first coding agent; the empire's daily driver for code work.

Free tier with limits. Pro: Rs 1,700/mo (USD 20). Business: Rs 3,400/seat/mo (USD 40).

Cursor, the IDE I Tried and the Empire's Soft Pass

Genuinely good IDE-pane AI; I just prefer the terminal flow for empire work.

Free, open source. Compute cost on consumer hardware is electricity, roughly Rs 4 to Rs 8 per active inference hour on a 65W desktop.

Ollama, the Local Model Runtime I Actually Trust

One binary, one model registry, zero cloud dependency. The default I reach for first.

Free self-hosted; Cloud tier from Rs 2,000/mo (lower run caps than Zapier at the same price)

n8n, Workflow Automation Without the Cloud Tax

Self-hostable, fair-code, and far cheaper than Zapier once you cross 1,000 runs a month.

Free open source; running cost depends on host (Rs 1,200/mo Oracle ARM via Coolify supports 5+ empire apps)

PocketBase, the Empire Backend I Run Across Every Project

SQLite plus auth plus realtime plus admin UI in one binary; I have not regretted picking it once.

Free open source; Oracle ARM free tier covers 4-core 24GB; Rs 1,200/mo for paid scale-out instances

Coolify on Oracle ARM, the Empire Hosting Stack

The Heroku-style PaaS I actually run, on the cheapest serious cloud bare-metal in 2026.

Free open source; running cost depends on host (Rs 1,200/mo Oracle ARM via Coolify is empire baseline)

FastAPI, the Empire's Default Python Serving Layer

Async by default, OpenAPI for free, and the framework I reach for first when I need an HTTP endpoint.

Today’s Sheet

Six more verdicts.

All sheets →
OPERATOR READ6/6

GLM-5.2 Cleared the Six Hard Tasks I Use to Vet Any Cheap Model

A new open-weights model matched my flagship on objective hard tasks. The battery I run before trusting any cheap model in production did not change.

  • A leaderboard rank is a reason to test a model, not a reason to trust it. I keep a fixed battery of objective hard tasks with execution-checked answers and run it on every cheap or open release bef…
  • On my June battery, GLM-5.2 matched Claude Opus 4.8 on all six tasks, from discrete optimization with a proof to a money-rounding fix.
ADITYA SHARMA · JUN 28, 2026 · 7 MIN
OPERATOR READ753B

GLM-5.2 Ships 753B Open Weights. My GTX 1660 Holds 6 GB.

The most powerful open-weight model of mid-June 2026 needs a cluster. A 4B model on a 6 GB card taught me what that headline leaves out.

  • Open weights and a model you can run are two different claims. GLM-5.2 ships MIT weights at over 750 billion parameters, and none of that helps a 6 GB card.
  • On a GTX 1660 the only runtimes that load are GGUF, CTranslate2 int8, and ONNX. No tensor cores means vLLM, FP8, and Marlin are off the table before you start.
ADITYA SHARMA · JUN 28, 2026 · 7 MIN
OPERATOR READ5

Claude Code v2.1.172 Unlocks Recursive Sub-Agents. My Fleet Found Three Walls.

Recursive sub-agents are a real upgrade, and after weeks of running CLI agent fleets in tmux, I can tell you exactly where the orchestration breaks.

  • Recursion is real, but it is not magic. v2.1.172 lets a sub-agent fan out its own sub-agents five levels deep, which means level two can quietly multiply your concurrency and your bill at the same …
  • Cold prompt-cache is the first wall. Fire N agents cold at once and every one misses the shared prefix. Fire one, wait for its stream, then fan out, and the rest read the cache at roughly a tenth o…
ADITYA SHARMA · JUN 28, 2026 · 9 MIN
DOSSIER-60-80%

Information Gain Is the 2026 Ranking Lever. Generic AI Content Is Suppressed.

The winning move is not more pages or more keywords. It is one genuinely new, first-hand thing per page. Everything else gets quietly shelved.

  • Information Gain is the central 2026 signal. Google's patent, granted in 2024, scores a page on what it adds that the rest of the results for that query do not already have. A reword of the top ten…
  • The reported impact is one-directional. Proprietary first-hand data pages up 15 to 25 percent, templated rewrites down 30 to 50 percent, AI-farm content down 60 to 80 percent. Sites publishing at s…
ADITYA SHARMA · JUN 1, 2026 · 7 MIN
OPERATOR READ90%

I Burned 90% Of GitHub's Free CI Minutes. Here's The Escape.

A real multi-repo empire eats 2000 free Actions minutes a month. When you hit zero, deploys stop firing silently. The fix is not paying per minute.

  • A multi-repo solo operator will exhaust 2000 free Actions minutes a month, not might, will. I hit 1800 of 2000 across one account with three active content repos, no macOS or Windows multiplier, ju…
  • If you set a zero-dollar spending budget, the cleanest setting, then at the limit your workflows hard-block until the first of the month. That means deploys silently stop firing and production goes…
ADITYA SHARMA · JUN 1, 2026 · 7 MIN
OPERATOR READ85%

Bing Sends One Of My Sites 85% Of Its Traffic. Google Sends 0.2%.

Every Indian operator I know optimises for Google. For a real audience segment, the traffic is on Bing, and Bing's index now feeds ChatGPT Search.

  • On one jobs-and-exam site I run, the Bing network is roughly 85 percent of organic traffic and Google organic is four sessions a week, 0.2 percent. The audience, govt-exam and job aspirants on Wind…
  • Yahoo India, DuckDuckGo, Ecosia, Copilot and ChatGPT Search all read off Bing's index. So the same Bing crawl that gives me 35 percent direct Bing clicks quietly powers another 40 percent across th…
ADITYA SHARMA · JUN 1, 2026 · 7 MIN
Deal Flow

Money, named losers, big numbers.

WHERE THE CAPITAL MOVED, WHO'S OUT, WHO'S NEXT.

More deals →
GLM-5.2

GLM-5.2 Ships 753B Open Weights. My GTX 1660 Holds 6 GB.

ADITYA SHARMA · 7 MIN

vibe coding

48 Hours to a Paying SaaS: Vibe Coding Tools Tested

AUTOKAAM EDITORIAL · 8 MIN

AI coding

Cursor vs Claude Code vs Zed vs Windsurf, May 2026 Field Test

AUTOKAAM EDITORIAL · 8 MIN

Anthropic

Anthropic Managed Agents, Nine Days In, The Operator Read

AUTOKAAM EDITORIAL · 7 MIN

AI policy

Commerce Cracks Frontier Models, Crowds Out Vendor Self-Audit

ADITYA SHARMA · 7 MIN

AI policy

Commerce Locks Google, Microsoft, xAI Into Pre-Release Federal Audits

ADITYA SHARMA · 7 MIN

Launch Wire

Shipped this week.

All launches →
FIELD NOTE20,000

Heretic 1.3 Ships Reproducible Runs, Bleeds the Fork-Cloners

Same abliteration tool, but now every published model carries a byte-for-byte recipe, and the mystique-merchants lose their cover.

  • Reproducible runs are the real release; benchmarking and VRAM are table stakes that should have shipped last year.
AUTOKAAM EDITORIAL · MAY 5, 2026 · 7 MIN
OPERATOR READ13M

Behind Monzo's Madrid Push, the US Retreat Loomed

Same neobank playbook, but Santander's Openbank already owns the Spanish SMB account before Monzo prints its first card.

  • Monzo lands Spain with 50 staff and zero products beyond a listening tour. Read it as a foundation move, not a market push.
ADITYA SHARMA · MAY 5, 2026 · 6 MIN
OPERATOR READ$2B

Moritz Raises $9M, Bleeds Harvey and Legora

The bet isn't selling AI to law firms. It's eating the law firm itself, $2B in contracts at a time.

  • Moritz isn't a vendor. It's a law firm that closed $2B in contracts with seven staff and a four-hour turnaround.
ADITYA SHARMA · MAY 5, 2026 · 7 MIN
SHEET88%

Hugging Face Guts Legacy OCR, Bleeds AWS Textract

Same model repo, new name on the box, but your inference pipeline just got a hard floor on version pinning.

  • v5.6.0 ships a tunable, on-prem PII filter that undercuts cloud-based Textract and Azure Form Recognizer, but your lockfile now breaks if you’re not pinned to latest.
AUTOKAAM EDITORIAL · APR 29, 2026 · 7 MIN
FIELD NOTE67.5k

MetaGPT Guts Lock-In, Bleeds Closed-Source Agent Stacks

Same notebook, new interpreter, but every indie dev now has a state-of-the-art autonomous agent that doesn’t answer to an API bill.

  • MetaGPT v0.8.0 ships a self-contained agent core that bypasses API-dependent stacks, a direct threat to closed-agent vendors charging per inference.
AUTOKAAM EDITORIAL · APR 29, 2026 · 6 MIN
FIELD NOTE$0/month

$0/month Over Vercel

Same production stack, but Oracle’s ARM instances made indie hosting free, and suddenly every side-project budget has room for PocketBase.

  • Oracle’s forever-free ARM instances (4 cores, 24GB RAM) are now the stealth GPU-tier for AI-native side projects, no billing dashboard, no surprise invoices.
AUTOKAAM EDITORIAL · APR 28, 2026 · 7 MIN
Hands-on

Tutorials & Playbooks.

All tutorials →
Claude Code Setup in India, Install, Pricing (INR), and Your First Project, claude on AutoKaam
ClaudeLevel / intermediate

Claude Code in India: How I Actually Run It Daily (Max 20x, Fork Subagents, the 5-Hour Meter)

I run Claude Code every working day across a dozen-plus projects from one Linux box in India. Here is the real Max 20x economics for a solo operator, the one env flag that fits 2x more agents in the same cap, /fast mode on Opus, and the 5-hour meter that punishes too many headless runs.

Published 8 min read
Build a WhatsApp AI Chatbot in 15 Minutes, For Small Businesses, business on AutoKaam
For BusinessLevel / beginner

WhatsApp AI Chatbot in 15 Minutes: The Build Is Easy, The Bans Are Not

A WhatsApp AI chatbot is genuinely a 15-minute build. Staying un-banned is the hard part. Here is the working Twilio plus AI stack, the compliance rule that nobody writes down, and the Baileys traps that ate my weekends.

Published 9 min read
Terminal showing a structuredData.json table extraction from a scanned PDF via Adobe PDF Services REST
AutomationLevel / intermediate

Programmatic PDF Table Extraction and OCR with Adobe PDF Services REST: The Auth, the Extract Call, and Parsing the Output

Published 8 min read
An AT-SPI2 accessibility tree of a GTK dialog with element names and roles, next to the same dialog being driven by an agent
AutomationLevel / advanced

I Gave My AI Agent Eyes and Hands on Native Linux Apps With AT-SPI2

Published 9 min read
Cloudflare named tunnel exposing a self-hosted app, kept reboot-proof with a systemd unit
AutomationLevel / intermediate

Reboot-Proof Cloudflare Named Tunnels: The systemd Setup I Run in Production

Published 8 min read
Directory

AI tools, compared.

Browse the directory →
✍️
Writing & Content
2 tools
🎨
Image Generation
5 tools
🎬
Video & Animation
6 tools
💻
Code & Development
8 tools
💬
Chat & Assistants
5 tools
🎵
Audio & Music
5 tools
🔍
Search & Research
3 tools
⚡
Productivity
5 tools
🎯
Design & UI
4 tools
📈
Marketing & SEO
5 tools

OpenAI

ChatGPT

Freemium

The AI assistant that started it all

Monthly
$20/mo
Rating
4.5 / 5

Anthropic

Claude

Freemium

The thoughtful AI assistant by Anthropic

Monthly
$20/mo
Rating
4.7 / 5

Midjourney Inc

Midjourney

Paid

The gold standard for AI art

Monthly
$10/mo
Rating
4.8 / 5

GitHub (Microsoft)

GitHub Copilot

Freemium

Your AI pair programmer

Monthly
$10/mo
Rating
4.3 / 5

Anysphere

Cursor

Freemium

The AI-first code editor

Monthly
$20/mo
Rating
4.5 / 5

Runway

Runway ML

Freemium

AI video generation and editing studio

Monthly
$15/mo
Rating
4.4 / 5
Operator playbook

Installed and tested. Real operator notes.

Claude: when to pick it over GPT→Cursor: IDE-pane vs terminal flow→Claude Code: the empire daily driver→ChatGPT: where it still wins→Gemini: real-time search edge→DeepSeek: cost-first API routing→

See all 30 operator playbook entries →

About

AutoKaam, operator's verdict on global SMB AI.

We don’t break news. We tell you what today’s break means for your stack. Every story is a verdict, one source quote, four operator bullets, one number, one named stake.

Coverage spans the frontier labs (Anthropic, OpenAI, Google DeepMind), agent platforms (Sierra, Cohere), developer tools (Cursor, Claude Code, Copilot), and the open-source frontier, for fifty-to-five-hundred-person firms in the US, UK, EU, Canada, and Australia.

Subscribe via RSS · browse by topic or author.