Skip to content

v0.2.1 - Complete Release

Choose a tag to compare

View all tags
@DevAnimecx DevAnimecx released this 03 Sep 13:37
· 31 commits to main since this release
v0.2.1
2b07ddb

Jiro v0.2.1 - Complete Release Notes

Local-first, AI-native web search & scraping platform with 12 social platforms, hybrid search, structured extraction, and Pro tier.

What's New

Phase 1: Search Intelligence

  • Hybrid Search - Combines keyword, semantic, and freshness signals for enhanced relevance
  • Cross-Encoder Reranking - ML-powered result reranking with configurable models
  • Semantic Embeddings - Vector similarity search using sentence-transformers
  • Relevance Scoring - Multi-signal ranking with keyword match, source authority, and freshness
  • Search Filters - Domain include/exclude, time range, category filters
  • Highlight Extraction - Query-aware snippet highlighting
  • Answer Synthesis - Extractive answers from search results
  • Multi-Query Expansion - Parallel query generation for complex topics

Phase 2: Social Scraping (12 Platforms)

Platform Features
Reddit Posts, comments, subreddits, user profiles
Hacker News Stories, comments, user profiles
YouTube Videos, channels, playlists, comments
Bluesky Posts, profiles, feeds
Twitter/X Tweets, threads, profiles (3-backend fallback)
Threads Posts, profiles
Instagram Posts, stories, reels, profiles
TikTok Videos, profiles, trending
LinkedIn Posts, profiles, companies
Facebook Posts, profiles, groups
Telegram Messages, channels, groups
Pinterest Pins, boards, profiles

Phase 3: Advanced Features

  • Structured Extraction - JSON Schema-based data extraction from search results
  • Intent Classification - Rule-based classifier with 16 intent types:
    • Informational, Navigational, Transactional, Research
    • Social, Shopping, News, Video, Image, Code
    • Academic, Local, Comparison, Definition, Troubleshooting, Creative
  • Smart Search - Auto-routing based on detected intent
  • Plugin System - 5 plugin types:
    • Engine Plugins (6): Google Scholar, arXiv, GitHub, Wikipedia, HN, Reddit
    • Search Plugins (6): Reranker, Deduplicator, Domain Filter, Freshness Boost, Source Authority, Snippet Enricher
    • Datasource Plugins (3): SEC Filings, Clinical Trials, Patents
    • Extractor Plugins: Custom content extraction
    • Social Plugins: Custom platform scrapers

Phase 4: Pro Tier

Plan Price RPM RPD Features
Free $0 10 100 Basic search, social scraping
Starter $29/mo 60 5,000 Hybrid search, structured extraction
Pro $99/mo 300 50,000 All features, webhooks
Enterprise $499/mo 1,000 500,000 Custom models, priority support
  • API Key Authentication - Secure key-based access with tiered permissions
  • Rate Limiting - Token bucket per API key
  • Quota Management - Daily request limits per tier
  • Usage Analytics - Detailed per-key, per-endpoint tracking

Phase 5: Production Ready

  • Docker - One-command deployment with docker-compose
  • Kubernetes - Production-ready Helm chart
  • OpenAPI 3.1 - Complete API specification
  • SDK Generation - Python, TypeScript, Go clients
  • Web Dashboard - Alpine.js + Tailwind UI with 5 tabs

Security Hardening

  • Auth enabled by default
  • Header-only API key authentication
  • Sanitized error messages
  • CORS protection
  • SSRF validation
  • JWT secret validation

API Endpoints

Search

  • POST /v1/search - Web search with hybrid mode
  • POST /v1/search/multi - Multi-query parallel search
  • GET /v1/search/stream - SSE streaming results

Social

  • POST /v1/social - Scrape social media URL
  • POST /v1/social/batch - Batch scrape multiple URLs
  • GET /v1/social/platforms - List supported platforms
  • POST /v1/social/search - Search across platforms
  • POST /v1/social/search/everywhere - Search all platforms

Smart

  • POST /v1/smart - Intent-aware smart search
  • POST /v1/smart/classify - Classify search intent

Structured

  • POST /v1/structured/extract - Extract structured data

Pro

  • POST /v1/pro/keys - Create API key
  • GET /v1/pro/keys - List API keys
  • DELETE /v1/pro/keys/{id} - Revoke API key
  • PUT /v1/pro/keys/{id}/upgrade - Upgrade tier
  • GET /v1/pro/usage - Get usage stats
  • GET /v1/pro/plans - List available plans

Monitor

  • GET /v1/monitor/status - Server status
  • GET /v1/monitor/health - Health check

MCP Tools (12)

Tool Description
search Search via 9 engines
scrape Extract URL content
ai_search Research with citations
search_hybrid Hybrid multi-signal search
search_structured Extract with JSON schema
social_scrape Scrape social URLs
social_search Search social platforms
social_batch Batch scrape URLs
smart_search Intent-aware routing
smart_classify Classify intent
compare_engines Compare engine results
monitor_status Server health metrics

Installation

pip

pip install jirosearch==0.2.1

npm (CLI wrapper)

npm install -g jiro-cli

Docker

docker pull jiro/jiro:0.2.1
docker run -p 8000:8000 jiro/jiro:0.2.1

Docker Compose

git clone https://github.com/DevAnimecx/jiro.git
cd jiro-search
docker-compose up -d

Kubernetes (Helm)

helm repo add jiro https://charts.jiro.dev
helm install jiro jiro/jiro --version 0.2.1

From Source

git clone https://github.com/DevAnimecx/jiro.git
cd jiro-search
pip install -e .

Quick Start

Start Server

jiro serve

Start Dashboard

jiro dashboard  # Opens on port 3000

Search

curl -X POST http://localhost:8000/v1/search \
  -H "Content-Type: application/json" \
  -d '{"q": "python web scraping", "engine": "google"}'

Hybrid Search

curl -X POST http://localhost:8000/v1/search \
  -H "Content-Type: application/json" \
  -d '{"q": "latest AI research", "hybrid": true, "answer": true}'

Social Scraping

curl -X POST http://localhost:8000/v1/social \
  -H "Content-Type: application/json" \
  -d '{"url": "https://reddit.com/r/programming/comments/abc123"}'

Smart Search

curl -X POST http://localhost:8000/v1/smart \
  -H "Content-Type: application/json" \
  -d '{"query": "github.com/fastapi"}'

Configuration

Environment Variables

# Server
JIRO_HOST=0.0.0.0
JIRO_PORT=8000

# Database
JIRO_DB_PATH=~/.jiro/jiro.db
JIRO_DB_POSTGRES=postgresql://user:pass@localhost/jiro

# Cache
JIRO_CACHE_TYPE=sqlite  # sqlite | memory | redis
JIRO_CACHE_TTL=3600

# Auth
JIRO_AUTH_ENABLED=true
JIRO_JWT_SECRET=your-secret-key-here

# Engines
JIRO_ENGINES=google,bing,brave,duckduckgo

Config File (~/.jiro/config.yaml)

server:
  host: 0.0.0.0
  port: 8000

engines:
  - google
  - bing
  - brave
  - duckduckgo

cache:
  type: sqlite
  ttl: 3600

auth:
  enabled: true
  jwt_secret: ${JIRO_JWT_SECRET}

MCP Integration

Claude Desktop

{
  "mcpServers": {
    "jiro": {
      "command": "jiro",
      "args": ["mcp"]
    }
  }
}

Cursor

{
  "mcpServers": {
    "jiro": {
      "command": "jiro",
      "args": ["mcp"]
    }
  }
}

Testing

# Run all tests
pytest tests/ -v

# Run specific test suite
pytest tests/test_phase1_search_intelligence.py -v
pytest tests/test_social_scraping.py -v
pytest tests/test_pro_tier.py -v

Security

  • Auth enabled by default
  • Header-only API key authentication
  • Sanitized error messages
  • CORS protection
  • SSRF validation
  • JWT secret validation

See SECURITY.md for reporting vulnerabilities.

Changelog

v0.2.1 (2026-09-03)

  • Security patch with auth hardening
  • Fixed CORS wildcard default
  • Removed hardcoded DB credentials
  • Sanitized error messages
  • Header-only API authentication

v0.2.0 (2026-09-03)

  • Initial v0.2 release with all features above

License

MIT License - see LICENSE for details.

Support

Credits

Built with ❤️ by the Jiro team.

Assets 6
Loading