DALAL
Sign in →
DALAL
Sign in →
Half of my pre-trade prep used to be asking the same setup to ChatGPT, then Claude, then Gemini, comparing tabs, second-guessing whichever sounded most confident on Bank Nifty levels or rupee positioning.
So I built Dalal — one prompt, six open-weight LLMs answering in parallel on Cloudflare's edge, with a synthesizer pointing out where they agreed and where one probably hallucinated. The cross-check before any trade you can't undo.
— Rajesh Kumar · Mumbai, India · Certified Financial Analyst
A single LLM will confidently support whichever side of the trade you frame your question around. Most retail traders use it as an amplifier of what they already wanted to hear.
What I do: Force six different models to answer in parallel, then read the disagreement. The disagreement is the signal.
Ask any single model for "Nifty 50 EMA 200 last close" and watch how often the number is wrong by 50+ points. Free, paid, big lab — all of them. Cross-checking is not optional.
What I do: Run the same factual query through six brains. If even one disagrees on the number, the number is suspect. I never trade on a single source.
Two hours of tab-switching between TradingView, ChatGPT, Twitter, and a broker terminal. Most of it spent rationalising a setup you already decided on at 8:45am.
What I do: One prompt to Dalal at 9am. Six brains, one verdict, agreement-vs-disagreement breakdown. Pre-trade prep collapses to 4 minutes.
Anyone can list brokerage rates and account minimums. The valuable signal is: which broker did real traders actually keep using after their first margin call?
What I do: I publish the live list of what I pay for monthly. When I cancel something, I write why.
Real agentic workflows need structured handoffs, retry logic, observability, and a way to not nuke your budget when one step loops. Most "agent frameworks" skip 3 of those 4.
What I do: Boring deterministic glue first. AI calls only where they earn their keep. Loud failure better than silent hallucination.
Generating 1000 articles with GPT and praying Google ranks them is a 2023 strategy that died in March 2024. Helpful Content Update killed it.
What I do: Process matters more than volume. Information gain per article. Specificity that doesn't generalize. Editorial voice the model can't fake.
Notion + Linear + Slack + Figma + 8 others = $400/mo and still confusion. Pick the 3 that match your actual workflow, kill the rest.
What I do: Cursor (or Claude Code) + Linear + 1 voice channel. Everything else is a distraction tax.
Best-in-benchmark tools die in production because the friction doesn't fit the workflow. The boring 80% solution that fits your habits beats the 100% solution you avoid.
What I do: Test for 1 week minimum, real work, real stakes. If I'm avoiding it by day 3, it's out.
Each analyst persona reflects a real workflow Indian retail traders face. The six LLMs are tuned around how these archetypes think — so the cross-check feels native to the problem.
Personas are illustrations of the analytical perspectives the six-LLM consensus engine is tuned around. Built independently · No broker partnerships
Ask one question. 6 open-weight LLMs answer in parallel on Cloudflare's edge — Llama 70B, DeepSeek R1, Qwen 2.5, Gemma 3, Mistral, Llama 8B. A judge model summarizes the consensus and scores agreement. The cross-check before any decision you can't undo.
For decisions one snapshot can't crack. 4 agents with opposing roles — Pessimist, Optimist, Engineer, Strategist — debate your dilemma across 3 rounds, rebut each other live, and a synthesizer closes the verdict. You watch it stream. The boardroom you don't have.
2-3x per week, my operator-voice take on what shifted in AI infra and tooling. No "10 best tools" listicles. Specific incidents, specific numbers, what to do about each. The narration alongside the machine.
