intelwatch 1.3.0 → 1.5.0

package/CHANGELOG.md

@@ -1,3 +1,13 @@
+## [1.3.2] - 2026-03-21
+### Added
+- **Google Gemini Provider**: Full support for Gemini models via Google API (`GEMINI_API_KEY` or `GOOGLE_API_KEY`) for Due Diligence AI analysis
+
+### Fixed
+- **BODACC Limit**: Reduced BODACC publication injection from 50 to 5 entries for AI context to drastically reduce token usage and prevent truncation
+- **M&A Output**: Reduced AI verbosity on health scores by enforcing strict bullet points constraints
+- **MaxTokens Override**: Removed hardcoded `maxTokens=1000` default and bumped to `8192` to allow full M&A history to generate without arbitrary cutoff
+- **Group Structure Render**: Fixed logic that accidentally injected Private Equity shareholders (like BPIFrance) into the subsidiaries array
+
 # CHANGELOG - v1.2 Draft
 
 ## Version 1.2.0 (en développement)
@@ -182,3 +192,8 @@ All notable changes to this project will be documented in this file.
 - M&A History PDF generation bug: Stopped truncating AI timeline events, full timeline is now preserved.
 - Group Structure Classification: Prevented PE Funds (BPIFrance, IK Partners, etc.) from being improperly categorized as operational subsidiaries in the AI due diligence report.
 - Fixed `pdfData` passthrough bug that caused empty PDF exports.
+
+### v1.3.1 (2026-03-21)
+- **Bug Fix**: Fixed an issue where PDF exports could contain empty Group Structure and M&A History sections leading to ugly page breaks.
+- **Bug Fix**: Preserved AI-generated subsidiaries when registry fallback has missing revenue data.
+- **Bug Fix**: Fixed PDF context scope throwing an undefined error when exporting via the `--export pdf` flag instead of the legacy `--format` option.

package/README.md

@@ -1,174 +1,128 @@
 # intelwatch
 
-> Competitive intelligence from the terminal. Track competitors, keywords, and brand mentions — no expensive SaaS required.
+> **Zero friction. Full context.**
+> Competitive intelligence, M&A due diligence, and OSINT directly from your terminal.
 
-## Install
+![Intelwatch Mockup](https://recognity.odoo.com/web/image/website/1/logo?unique=1)
+
+**Intelwatch** bridges the gap between hacker OSINT and B2B Sales/M&A data. It executes complex financial data aggregation, technology stack detection, and AI-powered due diligence in seconds. No dashboards, no bloated UI. Just pure data.
+
+## Installation
 
 ```bash
 npm install -g intelwatch
-# or from source:
-npm link
+# or run directly without installing:
+npx intelwatch profile kpmg.fr --ai
 ```
 
 **Requirements:** Node.js >=18
 
-## Quick Start
+## 🚀 The Flagship Feature: Due Diligence
 
-```bash
-# Add trackers
-intelwatch track competitor https://competitor.com --name "Acme Corp"
-intelwatch track keyword "audit SEO"
-intelwatch track brand "Recognity"
+Generate a comprehensive M&A/PE due diligence report in seconds. Intelwatch uses Smart Routing to fetch the best data based on the company's location (Pappers for France, Apollo/Clearbit/OpenCorporates for International).
 
-# Run checks
-intelwatch check
+```bash
+# Generate a deep profile with AI Due Diligence
+intelwatch profile doctolib.fr --ai
 
-# See what changed
-intelwatch digest
+# Export directly to a premium PDF report (Pro feature)
+intelwatch profile kpmg.fr --ai --export pdf
 
-# Full report
-intelwatch report --format html
+# Export data to Excel or CSV
+intelwatch profile "Acme Corp" --export xls
 ```
 
-## Commands
+**What it extracts:**
+- **Financials:** Revenue, net result, consolidated data, and growth analysis.
+- **Group Structure:** Shareholders, PE sponsors, and subsidiaries tree.
+- **Governance & M&A:** Board members, BODACC legal publications, and historical M&A timeline.
+- **OSINT & Reputation:** Recent press mentions, Reddit/Hacker News discussions, and sentiment scoring.
+- **Tech Stack:** Detection of 50+ technologies (CMS, Frameworks, Analytics, CDNs).
+- **AI Analysis:** Executive summary, Strengths & Weaknesses, Competitor identification, and Risk flags.
+
+---
 
-### Tracker Management
+## 🛠️ Core Commands
 
-#### `intelwatch track competitor <url> [--name <alias>]`
-Tracks a competitor website. Captures:
-- Pages found (via link extraction)
-- Pricing page content and price changes
-- Technology stack (35+ technologies detected)
-- Open job positions (/careers, /jobs)
-- Social links
-- Meta title/description changes on key pages
+### 1. Market & Competitor Discovery
+Discover actual competitors for any website using web search and AI scoring.
 
 ```bash
-intelwatch track competitor https://acme.com --name "Acme"
-intelwatch track competitor https://rival.io
+intelwatch discover https://mycompany.com --export csv
 ```
 
-#### `intelwatch track keyword <keyword> [--engine google]`
-Tracks Google SERP rankings for a keyword. Records top 20 results, detects position changes, new entrants/exits, and featured snippet holders.
+### 2. Track Competitors & Keywords
+Set up local trackers to monitor competitor websites, technology changes, and Google SERP rankings over time.
 
 ```bash
-intelwatch track keyword "project management software"
+intelwatch track competitor https://acme.com --name "Acme Corp"
 intelwatch track keyword "audit SEO" --engine google
+intelwatch track brand "Recognity"
 ```
 
-#### `intelwatch track brand <name>`
-Tracks brand mentions across Google News and recent web results. Detects sentiment (positive/negative) and categorizes mentions (press, blog, forum, social, review).
+### 3. Check & Digest
+Run your trackers and see what changed since the last snapshot.
 
 ```bash
-intelwatch track brand "Recognity"
-intelwatch track brand "My Company Name"
+intelwatch check
+intelwatch digest
+intelwatch diff acme-com --days 7
 ```
 
-### Listing & Removing
+### 4. AI Briefs & Sales Pitches
+Generate AI-powered competitive briefs and sales pitches against your tracked competitors.
 
 ```bash
-intelwatch list                    # List all trackers
-intelwatch remove <tracker-id>     # Remove a tracker
+intelwatch ai-summary
+intelwatch pitch acme-com
 ```
 
-### Checking & Diffs
+---
 
-```bash
-intelwatch check                   # Check all trackers
-intelwatch check --tracker acme-com  # Check one tracker
+## 💎 Pro License ($49/mo)
 
-intelwatch diff acme-com           # Compare last 2 snapshots
-intelwatch diff acme-com --days 7  # Compare with 7 days ago
-```
+Intelwatch operates on a freemium model. The **Free Tier** allows standard OSINT, basic company profiles, and CSV exports.
 
-### Reports
+The **Pro Tier** unlocks:
+- 📄 **Premium PDF & Excel (XLS) exports**
+- 🤖 **AI Due Diligence Reports** (Health score, Risks, M&A timelines)
+- 🌍 **International Routing** (Apollo & Clearbit integrations)
+- 🕵️ **Deep OSINT** (Reddit & Hacker News tracking)
 
+**Activate your license:**
 ```bash
-intelwatch digest                        # Quick summary table
-intelwatch report                        # Markdown report (stdout)
-intelwatch report --format html          # HTML report (saved to ~/.intelwatch/reports/)
-intelwatch report --format json          # JSON report (stdout)
-intelwatch report --format md --output ./weekly.md  # Custom output file
+intelwatch auth YOUR_LICENSE_KEY
 ```
+*(Get your key at [recognity.fr/tools/intelwatch](https://recognity.fr/tools/intelwatch))*
 
-### History & Comparison
-
-```bash
-intelwatch history acme-com              # Show snapshot history
-intelwatch history acme-com --limit 10  # Last 10 snapshots
+---
 
-intelwatch compare acme-com rival-com    # Side-by-side comparison
-```
+## ⚙️ Configuration & API Keys
 
-### Notifications
+Intelwatch brings your own keys (BYOK) for maximum privacy and limit-less scaling. Set these in your environment variables (`~/.bashrc` or `~/.zshrc`):
 
 ```bash
-intelwatch notify --setup    # Interactive setup (Slack, Discord webhook)
-```
-
-Notifications config is stored at `~/.intelwatch/config.yml`:
-
-```yaml
-notifications:
-  webhook: https://hooks.slack.com/services/xxx/yyy/zzz
-  events:
-    - competitor.new_page
-    - competitor.price_change
-    - keyword.position_change
-    - brand.new_mention
-    - brand.negative_mention
-```
+# Required for AI Features (Choose one)
+export ANTHROPIC_API_KEY="sk-ant-..."
+export OPENAI_API_KEY="sk-proj-..."
 
-## Data Storage
+# Required for Web Search & Discovery
+export BRAVE_SEARCH_API_KEY="BSAeG..."
 
-All data is stored locally in `~/.intelwatch/`:
+# Required for deep French corporate data
+export PAPPERS_API_KEY="86c0dcc..."
 
+# Required for deep International corporate data (Pro)
+export APOLLO_API_KEY="..."
 ```
-~/.intelwatch/
-├── config.yml           # Notification settings
-├── trackers.json        # Active trackers
-├── snapshots/           # Historical snapshots (JSON)
-└── reports/             # Generated HTML reports
-```
-
-## Technology Detection
-
-Detects 35+ technologies via headers, meta tags, scripts, HTML patterns:
 
-| Category | Technologies |
-|----------|-------------|
-| CMS | WordPress, Drupal, Joomla |
-| E-commerce | Shopify, Magento |
-| Website Builder | Wix, Squarespace, Webflow |
-| JS Framework | React, Vue.js, Angular, Next.js, Nuxt.js, Gatsby, Svelte |
-| JS Library | jQuery |
-| CSS Framework | Bootstrap, Tailwind CSS |
-| Analytics | Google Analytics, Google Tag Manager, Facebook Pixel, Hotjar |
-| CRM/Marketing | HubSpot, Mailchimp, Intercom |
-| CDN/Security | Cloudflare |
-| Web Server | nginx, Apache |
-| Hosting | Vercel, Netlify |
-| Backend | PHP, Django, Ruby on Rails, Node.js/Express |
-| Payment | Stripe |
+*(You can also use a `.env` file in your working directory).*
 
-## Sentiment Analysis
-
-English and French word lists for positive/negative detection in brand mentions. Categorizes mentions as: press, blog, forum, social, or review.
-
-## Design Principles
-
-- **No external APIs** — everything via respectful web scraping
-- **Respectful scraping** — 1-2s delays, user-agent rotation, retry backoff
-- **Graceful degradation** — saves what it can if a check partially fails
-- **Local-first** — all data stays on your machine
-
-## Tests
-
-```bash
-npm test
-```
+## 🔒 Privacy & Architecture
 
-40 tests covering storage logic, technology detection, and sentiment analysis.
+- **Local-first**: All tracker data, snapshots, and configurations are stored locally in `~/.intelwatch/`.
+- **No intermediary servers**: The CLI talks directly to the data providers (Pappers, Apollo, Brave, Anthropic/OpenAI). We do not see your API keys or your searches.
+- **Smart Routing**: `registry.js` automatically detects French companies (via SIREN/SIRET or country code) and routes them to Pappers, while international companies gracefully fallback to Apollo and OpenCorporates.
 
 ## License
 

package/ROADMAP-PREMIUM.md

@@ -1,6 +1,6 @@
 # Intelwatch Deep Profile — Roadmap Premium
 
-## Current (v1.1.x — Shipped)
+## Current (v1.3.x — Shipped)
 - [x] Identity + finances 5 ans + consolidé groupe
 - [x] Filiales via recherche-dirigeants + CA/résultat
 - [x] Représentants / actionnariat
@@ -22,25 +22,25 @@
 - [x] **Article scraping for M&A depth** — top 5 articles (2000 chars), content injected into AI
 - [x] **Key date extraction from articles** — regex code-side, authoritative dates
 - [x] **FLI code-built revenue target** — scans all articles, picks highest announced target, overrides AI
-- [x] **Stale year warning** — red ⚠️ badge on subsidiaries with data > 2 years old
+- [x] **Stale year warning** — red ⚠️ badge on subsidiaries avec data > 2 ans
+- [x] Export JSON/CSV/XLS structuré
+- [x] `--lang fr` option (PDF + AI prompts in French)
+- [x] **OpenCorporates** — filiales internationales
+- [x] **Smart Routing & International Pivot** — Apollo.io + Clearbit + OpenCorporates
+- [x] OSINT scraping (Reddit, HackerNews)
+- [x] Freemium gate (Free vs Pro) + license key check
+- [x] Fix forces/faiblesses vides en terminal (parsing JSON)
+- [x] Commit + npm publish as `intelwatch@1.3.x`
 
-## v1.2 — Next Release
-- [ ] Fix forces/faiblesses vides en terminal (parsing JSON)
-- [ ] Export JSON/CSV structuré
-- [ ] `--lang fr` option (PDF + AI prompts in French)
-- [ ] Comparable transactions section (competitors' M&A/fundraising with article links)
-- [ ] Geographic implantations scraping from company website
-- [ ] Cross-reference press/journalists across sections
-- [ ] Commit + npm publish as `intelwatch@1.2.0`
-
-## V2 — Pro ($49/mo)
+## V2 — Pro ($49/mo) — Next Releases
 - [ ] `intelwatch compare SIREN1 SIREN2` — côte à côte
 - [ ] **INPI integration** — brevets & marques du groupe (gratuit, data.inpi.fr)
-- [ ] **OpenCorporates** — filiales internationales (gratuit)
 - [ ] **Annuaire Entreprises / data.gouv** — données complémentaires
-- [ ] Freemium gate (3 profiles/jour) + license key
 - [ ] `--preview` mode limité (identity + dernier exercice)
 - [ ] **BODACC détaillé enrichi** — timeline types d'actes, augmentations capital
+- [ ] Comparable transactions section (competitors' M&A/fundraising with article links)
+- [ ] Geographic implantations scraping from company website
+- [ ] Cross-reference press/journalists across sections
 
 ## V3 — Deep Profile ($299/mo)
 - [ ] **Graphe de liens** — visualisation SVG/HTML du réseau (Pappers graph-style)
@@ -64,7 +64,7 @@
 | Pappers | ✅ (BYOK) | ✅ API | v1.0 ✅ |
 | Brave Search | ✅ (BYOK) | ✅ API | v1.0 ✅ |
 | INPI (brevets/marques) | ✅ | ✅ data.inpi.fr | V2 |
-| OpenCorporates | ✅ | ✅ | V2 |
+| OpenCorporates | ✅ | ✅ | v1.3 ✅ |
 | Annuaire Entreprises | ✅ | ✅ data.gouv | V2 |
 | BODACC détaillé | ✅ | ✅ bodacc.fr | V2 |
 | Infogreffe | ❌ 3.80€/doc | API payante | V4 |
@@ -77,8 +77,8 @@
 ## Pricing
 | Tier | Prix | Limites | Features |
 |------|------|---------|----------|
-| Free | 0€ | 3 profiles/jour, `--preview` | Identity + last year only |
-| Pro | 49€/mo | Illimité | Full profile + AI + PDF + export + INPI + compare |
+| Free | 0€ | Bypass limites CSV | Identity + last year only |
+| Pro | 49€/mo | Illimité | Full profile + AI + PDF + export XLS + INPI + Reddit/HN |
 | Deep Profile | 299€/mo | Illimité | Tout Pro + graphe + alertes + batch + scoring |
 | Enterprise | Custom | Custom | Tout Deep + Infogreffe + SEMrush + API REST |
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "intelwatch",
-  "version": "1.3.0",
+  "version": "1.5.0",
   "description": "Competitive intelligence CLI — track competitors, keywords, and brand mentions from the terminal",
   "type": "module",
   "license": "MIT",
@@ -8,7 +8,7 @@
     "intelwatch": "./bin/intelwatch.js"
   },
   "scripts": {
-    "test": "node --test test/storage.test.js test/sentiment.test.js test/tech-detect.test.js test/export.test.js test/i18n.test.js test/error-handler.test.js test/reddit-hn.test.js test/license.test.js test/providers.test.js",
+    "test": "node --test test/storage.test.js test/sentiment.test.js test/tech-detect.test.js test/export.test.js test/i18n.test.js test/error-handler.test.js test/reddit-hn.test.js test/license.test.js test/providers.test.js test/searxng-search.test.js test/compare.test.js",
     "start": "node bin/intelwatch.js"
   },
   "dependencies": {

package/src/ai/client.js

@@ -5,9 +5,13 @@ import { loadConfig } from '../config.js';
  * Returns null if no key is configured.
  */
 export function getAIConfig() {
+  const envGoogle = process.env.GOOGLE_API_KEY || process.env.GEMINI_API_KEY;
   const envOpenAI = process.env.OPENAI_API_KEY;
   const envAnthropic = process.env.ANTHROPIC_API_KEY;
 
+  if (envGoogle) {
+    return { provider: 'google', apiKey: envGoogle, model: 'gemini-2.5-flash' };
+  }
   if (envOpenAI) {
     return { provider: 'openai', apiKey: envOpenAI, model: 'gpt-4o-mini' };
   }
@@ -44,17 +48,28 @@ export function hasAIKey() {
  * Throws on API errors.
  */
 export async function callAI(systemPrompt, userPrompt, options = {}) {
+  const maxTokens = options.maxTokens || 1000;
+
+  if (options.uncensored) {
+    const host = process.env.OLLAMA_HOST || 'http://localhost:11434';
+    // We default to llama3 for uncensored local OSINT if the user hasn't specified one
+    const model = process.env.OLLAMA_MODEL || 'llama3';
+    return callOllama(host, model, systemPrompt, userPrompt, maxTokens);
+  }
+
   const aiConfig = getAIConfig();
   if (!aiConfig) {
     throw new Error(
-      'No AI API key configured. Set OPENAI_API_KEY or ANTHROPIC_API_KEY, ' +
-      'or add ai.api_key to ~/.intelwatch/config.yml'
+      'No AI API key configured. Set GEMINI_API_KEY, OPENAI_API_KEY or ANTHROPIC_API_KEY, ' +
+      'or add ai.api_key to ~/.intelwatch/config.yml. Use --uncensored for local Ollama.'
     );
   }
 
   const { provider, apiKey, model } = aiConfig;
-  const maxTokens = options.maxTokens || 1000;
 
+  if (provider === 'google') {
+    return callGoogle(apiKey, model, systemPrompt, userPrompt, maxTokens);
+  }
   if (provider === 'anthropic') {
     return callAnthropic(apiKey, model, systemPrompt, userPrompt, maxTokens);
   }
@@ -128,3 +143,67 @@ export function estimateCost(inputChars, outputChars, provider = 'openai') {
   const cost = inputTokens * rates.in + outputTokens * rates.out;
   return { inputTokens, outputTokens, cost: cost.toFixed(5) };
 }
+
+
+async function callGoogle(apiKey, model, systemPrompt, userPrompt, maxTokens) {
+  // Use v1beta for Gemini 2.5
+  const url = `https://generativelanguage.googleapis.com/v1beta/models/${model}:generateContent?key=${apiKey}`;
+  const res = await fetch(url, {
+    method: 'POST',
+    headers: {
+      'Content-Type': 'application/json',
+    },
+    body: JSON.stringify({
+      systemInstruction: { parts: [{ text: systemPrompt }] },
+      contents: [{ role: 'user', parts: [{ text: userPrompt }] }],
+      generationConfig: {
+        maxOutputTokens: maxTokens,
+        temperature: 0.2
+      }
+    }),
+  });
+
+  if (!res.ok) {
+    const body = await res.text();
+    throw new Error(`Google API ${res.status}: ${body.slice(0, 200)}`);
+  }
+
+  const data = await res.json();
+  if (!data.candidates || !data.candidates[0].content) {
+    throw new Error('Invalid Google API response');
+  }
+  return data.candidates[0].content.parts[0].text.trim();
+}
+
+async function callOllama(host, model, systemPrompt, userPrompt, maxTokens) {
+  const url = `${host.replace(/\/$/, '')}/api/chat`;
+  const res = await fetch(url, {
+    method: 'POST',
+    headers: {
+      'Content-Type': 'application/json',
+    },
+    body: JSON.stringify({
+      model: model,
+      messages: [
+        { role: 'system', content: systemPrompt },
+        { role: 'user', content: userPrompt }
+      ],
+      stream: false,
+      options: {
+        num_ctx: 16384,
+        num_predict: maxTokens
+      }
+    }),
+  });
+
+  if (!res.ok) {
+    const body = await res.text();
+    throw new Error(`Ollama API ${res.status}: ${body.slice(0, 200)}`);
+  }
+
+  const data = await res.json();
+  if (!data.message || !data.message.content) {
+    throw new Error('Invalid Ollama API response');
+  }
+  return data.message.content.trim();
+}