> For the complete documentation index, see [llms.txt](https://open-papertrade.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://open-papertrade.gitbook.io/docs/getting-started/first-run.md). # First Run You've installed the app and both servers are running. Here's a 5-minute path that touches every major feature. ## 1. Create your account Open → **Sign up** (top right). Verification email is a no-op in dev — you'll be logged in immediately. ## 2. Look around the dashboard The sidebar shows every feature: | Item | What it does | | -------------------- | ---------------------------------------------------- | | Dashboard | Portfolio overview, market movers, quick stats | | Markets | Live quotes, stock search, sector heatmap | | Charts | Interactive TradingView-style charts with indicators | | Portfolio | Your holdings + P/L | | Trade | Place market or limit orders | | History | Transaction log | | Watchlist | Track symbols without owning them | | Reports | Downloadable performance reports | | Backtesting | Build & test trading strategies over historical data | | AI Coach | Behavioral analysis of your trades | | **Filings Research** | Ask questions of SEC 10-K / 10-Q / 8-K filings | | Copy Trading | Follow leader accounts, mirror their trades | | Leaderboard | Rank by return, XP, achievements | | Friends | Social layer | | Account | Settings, 2FA, currency | ## 3. Make your first trade **Trade → Buy** — search for `AAPL`, buy 10 shares. You start with a virtual balance so nothing costs you anything. Once it fills, check: * **Portfolio** — the position appears with unrealized P/L. * **History** — the transaction is logged with fill price. ## 4. Add a filing to the Research corpus Sidebar → **Filings Research**. The corpus starts **empty** — that's on purpose. ### Option A — Paste a URL in the UI Grab any SEC filing URL. Easiest path: 1. Go to [sec.gov/cgi-bin/browse-edgar?action=getcompany](https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany). 2. Search for a company (e.g. "Apple"). 3. Click any 10-K filing → click the primary `.htm` document. 4. Copy the URL — it looks like `https://www.sec.gov/Archives/edgar/data/320193/000032019324000123/aapl-20240928.htm`. Paste it into the **"Add a filing by URL"** box → click **Ingest**. Watch the job status flip through `pending → running → success` in \~60 seconds. When done, the filing appears in the "Ingested filings" list below. ### Option B — Bulk seed via CLI If you want a bigger corpus fast, open a third terminal: ```bash cd backend source .venv/bin/activate python manage.py ingest_filings AAPL NVDA TSLA --form 10-K --limit 1 ``` This takes 5–10 minutes for three filings on CPU. ## 5. Ask your first question Back on **Filings Research**: * Type: **"What are the largest risk factors disclosed in this filing?"** * Leave **Hybrid** and **Rerank** toggles **on** (recommended defaults). * Click **Ask**. In 3–8 seconds you get an answer with inline `[S1]`, `[S2]`, `[S3]` citation markers. Below the answer, citation cards show: * Ticker, form type, date, section name. * A snippet of the quoted text. * A `↗ source` link that opens the exact filing on sec.gov. **Click any citation superscript in the answer** — it jumps you to the source URL. Every claim is verifiable in one click. ## 6. Try a refusal (this is the point) Ask: **"Did Apple acquire LithGold Corporation in 2024?"** The response should say: > **🚫 declined** — I don't have enough information in the provided filings to answer that. No hallucination. No polite fabrication. This is the differentiator over 90% of "chat with your PDF" tools. ## 7. Try Agent mode Ingest a second company (say, NVDA) if you haven't already, then check **Agentic mode** and ask: > **"Compare Apple's and NVIDIA's approach to R\&D investment."** Wait 8–15 seconds. The response includes: * A synthesized comparison covering both companies. * An **Agent trace** disclosure showing the sub-questions the planner generated (one per ticker). * Merged citations from both filings, with global numbering. ## 8. Peek at the eval harness ```bash python manage.py evaluate --hybrid --rerank --out results.json cat results.json ``` You'll see recall\@k, MRR, and answer-level metrics on the gold set. See [Evaluation Harness](/docs/filings-deep-dive/evals.md) for full methodology. *** ## What next?
Learn features in depthEvery feature has its own page with demos and walkthroughs.Overview
Understand the Filings pipelineHow ingestion, hybrid retrieval, and the agent actually work.Architecture
Something broke?Common issues + fixes.Troubleshooting