🔬 Strategy Lab

What Is Strategy Lab?

Sequential Agent Chains

Unlike a single AI prompt, Strategy Lab runs a chain of specialised AI analysts in sequence. Each agent reads the previous agent's complete output before writing its own. This means the Sector Analyst knows the full macro picture before picking sectors, and the Stock Selector reads both macro and sector analysis before naming stocks. The final report compiler synthesises everything into a structured brief.

Your Key, Your Cost

Strategy Lab uses your own BYO API key exclusively. Finmagine does not call any paid AI model on your behalf — zero additional billing from us. Groq and Gemini offer free API tiers that cover most runs at no cost. You add your key once in Account Settings, and Strategy Lab uses it for every run you initiate.

Setting Up Your BYO API Key

Supported Providers

Adding Your Key

1. Go to Account Settings → BYO API Keys.
2. Click Add Key next to your chosen provider and paste your API key.
3. Click Save. The key is encrypted with AES-256 before storage and decrypted only server-side when you initiate a run — it is never exposed in logs or responses.
4. Return to Strategy Lab — your key appears as a selectable card in Step 1.

The 3-Step Workflow

Every Strategy Lab run follows the same three steps. The Execute button stays disabled until all three are complete — preventing accidental runs and API calls.

Step 1 — Select Your AI Provider

The provider cards in Step 1 show each provider whose key you have saved. Click the card to select it — it highlights when active. You can save keys for multiple providers and switch between them per run. If no cards appear, add a key in Account Settings.

Step 2 — Choose a Research Preset

Four preset cards are shown. Click one to select your research team — it highlights when active. Each preset is a different agent composition covering a different analytical framework. The preset determines which agents run, in what order, and what each one covers.

Step 3 — Describe Your Research Goal

Type a specific research question or investment theme in the goal field. This text is injected into every agent's prompt — it is the thread that connects all the agents' outputs into a coherent report.

The 4 Research Presets

Each preset is a different analyst team with a different agent chain. Choose based on the type of research question you have.

📊 Equity Research Team — Agent Details

This is the flagship preset. It mirrors how a sell-side research team works: a macro analyst sets the scene, a sector analyst identifies where the opportunity lies, a stock selector picks specific names, and a report compiler synthesises everything into a structured brief.

🏢 Fundamental Research — Agent Details

Purpose-built for bottom-up investors who already know the sector they want to research and need a structured fundamental evaluation. Works well for questions like "assess the investment case for quality NBFC stocks" or "evaluate the fundamental attractiveness of Indian hospital chains."

🔢 Quant Strategy Desk — Agent Details

For systematic and factor-based investors. It does not pick individual stocks — it designs a factor strategy blueprint and produces AI-generated historical expectations to verify. The Strategy Validator agent discusses how the strategy would likely have behaved historically; it does not run a live backtest. Use it for questions like "design a momentum + quality factor screen for NSE midcaps" or "build a low-volatility dividend yield strategy for defensive positioning."

📈 Technical Analysis Panel — Agent Details

For traders and technically-oriented investors. Focuses on price structure, momentum, and setup identification. Works well for questions like "assess the IT sector technical picture heading into results season" or "identify FMCG technical setups for a potential mean-reversion trade."

Which Preset Should I Use?

Reading the Output & Run History

Output Structure

When a run completes, the final agent's output appears as a formatted Markdown report in the output panel. Each preset produces a consistent section structure — the Report Compiler or final agent always uses a clear heading hierarchy so you can scan to the section most relevant to you.

The stock picks or recommendations within any report are AI-generated from the model's training knowledge. They do not pull live Finmagine data (prices, screener results, ECS scores). Always cross-check specific names using Finmagine's Screener, ECS, or individual stock pages before acting on any idea.

Run History

Below the main run panel, the History section shows your last 20 Strategy Lab runs. Each row displays the preset used, your market and goal inputs, the run status, and the date. Click View Report on any completed run to reload the full output without re-running it. Reports are stored server-side and persist across sessions and devices.

If a run shows as Failed, the most common cause is a provider rate limit or a temporary API outage. Start a new run with the same settings — or switch to a different provider. Groq and Gemini free tiers rarely fail simultaneously.

Related Guides