24, Value proposition 2026: multi-market repositioning
Auteur : strategic review post-multi-market launch Date : 2026-05-16 Périmètre : repositionnement de la plateforme suite à l'extension multi-marchés (FR/US/DE/CH/UK + CA early), API publique, MCP server.
TL;DR, Positioning statement
Insiders Trades Sigma is the only open multi-market insider-trades intelligence layer. We unify AMF (France), SEC Form 4 (US), BaFin (Germany), SIX SER (Switzerland), RNS/Investegate (UK), and soon Canada, into one FX-normalised schema, backtest the resulting signals against local benchmarks, and expose the whole stack via a public REST API and an MCP server so AI agents can consume it natively. We are to insider data what OpenInsider was to SEC alone, ten years later, pan-European, backtest-augmented, AI-native.
Where competitors do one of these (OpenInsider = US-only free, Quiver = US alt-data SaaS, WhaleWisdom = 13F, Bloomberg = paywalled everything), we do the intersection: multi-market + normalised + backtested + open API + MCP.
Pyramid: who pays / who uses free / who promotes
| Layer | Persona | Behaviour | Monetisation |
|---|---|---|---|
| Promoters (top) | Retail finance influencer, fintwit account, financial journalist | Cite signal screenshots, embed our charts, reference our methodology | Free + attribution required |
| Free users (mass) | EU retail trader, hobbyist quant, student | Browse companies, insiders, top signals, basic backtest viewer | Free, rate-limited (60 req/min IP, no API key needed) |
| Paying, Developer | Indie quant, side-project builder, AI agent operator | Hits API daily, builds Telegram bots, GPTs, Notion dashboards | $19/mo, 10k req/mo, full schema, 1 webhook |
| Paying, Pro | Boutique fund analyst, family office, RIA | Pulls every filing into internal models, alerts on portfolio | $99/mo, 100k req/mo, unlimited webhooks, MCP access, CSV exports, multi-seat |
| Enterprise (later) | Hedge fund, asset manager, data vendor | White-label, SLA, dedicated schema | Custom, $500–$2k/mo |
The economic gravity is in the Developer tier: AI agents and GPT builders need cheap, well-structured insider data the same way they need stock prices. We aim to be the Polygon.io of insider data.
Competitive matrix
| Capability | Sigma | OpenInsider | Quiver Quant | WhaleWisdom | Fintel | Bloomberg | Capital IQ |
|---|---|---|---|---|---|---|---|
| US Form 4 | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
| EU insider filings (AMF/BaFin/SIX/RNS) | ✓ unified | ✗ | ✗ | ✗ | partial | ✓ paywalled | ✓ paywalled |
| FX-normalised cross-market | ✓ | ✗ | ✗ | ✗ | ✗ | ✓ | ✓ |
| Composite signal score | ✓ open formula | ✗ | ✓ proprietary | partial | ✓ proprietary | ✗ | ✗ |
| Backtest viewer (web) | ✓ free | ✗ | paywalled | paywalled | paywalled | paywalled | paywalled |
| Public REST API | ✓ free tier | ✗ | ✓ paid | ✗ | ✓ paid | ✓ enterprise | ✓ enterprise |
| MCP server (AI-native) | ✓ first | ✗ | ✗ | ✗ | ✗ | ✗ | ✗ |
| Pricing entry point | $0 / $19 | $0 | $10/mo | $50/mo | $30/mo | $24k/yr | $30k/yr |
| Open methodology | ✓ full | partial | ✗ | ✗ | ✗ | ✗ | ✗ |
Three callouts:
- vs OpenInsider, Sigma is what OpenInsider would look like in 2026 if it had been rebuilt for the EU, normalised currencies, added a backtest engine, and published an API. OpenInsider is US-only, free, no API, no scoring.
- vs Quiver Quant, Quiver is US-centric, proprietary score, paywalled API. Sigma is multi-market, open formula, free tier API, plus an MCP endpoint Quiver does not have.
- vs Bloomberg / CapIQ, they have everything but cost five figures a year and you cannot point an LLM at them. Sigma is 1% of the price for the specific insider-trades slice, with AI-agent access built in.
Pricing, 3 tiers
Free, Explorer ($0)
- Full web UI: companies, insiders, signals, methodology, backtest viewer
- 60 requests/minute per IP, no API key required
- 7-day signal history on API responses
- Attribution required for republication
Developer, $19/mo
- 10,000 API requests/month
- Full historical depth (2020 onwards, all 5 markets)
- 1 webhook destination
- MCP server (read-only)
- Email alerts on watched tickers
- CSV export (single ticker, ad hoc)
Pro, $99/mo
- 100,000 API requests/month
- Unlimited webhook destinations
- MCP server (full read + portfolio context)
- Bulk CSV / Parquet exports
- Multi-seat (up to 5)
- Portfolio attribution: matched signals against your holdings
- Priority email support (24h)
CTA buttons: Get API key (signup) for Free + Developer, Contact sales (mailto:simon.azoulay.pro@gmail.com?subject=Sigma%20Pro) for Pro.
6 use cases
- Retail discovery, A French retail trader browses /companies, filters by market = US, sees that an Atlassian SVP just bought $2.4M at a 30-day low. T+90 historical win rate on similar signals: 71%.
- Quant signal feed, A boutique fund pulls
/api/v1/signals?score_gte=70&market=eudaily into their factor model. Insider score becomes the 7th column next to value/momentum/quality. - AI agent context, A custom GPT or Claude agent connects to our MCP server, answers “did any Allianz director buy in the last 30 days?” in one tool call, with structured data.
- Newsroom alerts, A financial journalist sets a webhook on score ≥ 80 + market cap ≥ €10B. Gets pinged in Slack 5 minutes after a CEO files. Writes the story before competitors.
- ETF strategy, A fund builds a long-only basket of top-30 insider-score names per market, rebalanced monthly. Sigma's backtest viewer shows the rolling Sharpe of that basket without writing code.
- Compliance research, A regulator or academic researcher reproduces our backtest with the open formula, audits selection bias, cites the methodology page in a working paper.
5 narrative angles for marketing
- “Form 4 ↔ AMF ↔ BaFin: first time in one schema.” No one else has unified the four filing systems. We did it. Here is the schema, here is the open-source mapping table.
- “The Polygon.io of insider data.” Cheap, fast, well-documented, AI-friendly. The infra layer indie quants and AI agents wish existed.
- “OpenInsider for Europe, and then some.” A familiar reference point for the US audience. Same free spirit, broader scope, modern stack.
- “The first MCP server for insider filings.” Aimed at the AI-agent builder crowd. We ship the tool spec, examples for Claude Desktop, ChatGPT desktop, Cursor.
- “Backtest first, claim second.” Every score we display is backed by a public backtest with disclosed selection biases. We publish the misses, not just the hits.
Roadmap implications (out of scope for this memo, flagged)
- CA needs ≥12 months data before being elevated from “beta” badge to first-class market.
- JP (EDINET) blocked on JP phone requirement, workaround via local agent.
- API monetisation requires Stripe + API-key auth scaffolding (not yet shipped).
- MCP server is shipped read-only; portfolio-context tier requires user-auth wiring.
Document closes the strategic loop. Implementation files listed in PR description.