SkillTide Submit
Official description7 sourcesMCP servers8d ago
parastejpal987-cmyk/opticparse-public

parastejpal987-cmyk/opticparse-public

Stealth Multimodal Web Scraper & 0-Day Phishing Shield for AI Agents. Alternative to Firecrawl & Crawl4AI with zero-CSS vision extraction, 200 free trial credits, and native LangChain / ElizaOS / MCP support.
View on GitHub ↗Direct link · no redirect page
Config template (adjust for your client)
claude mcp add opticparse-public -- npx -y opticparse-public
A pasteable skeleton generated from the repository name. Check the project README for the exact package and arguments.
Star count
2 ★ · parastejpal987-cmyk/opticparse-public · read from the repository page · (checked 2026-10-16 20:00:20)
Source
collected from the punkpeye/awesome-mcp-servers list · type: MCP server
Harvest history
first seen 2026-10-09 · last seen 2026-10-16 · seen 32× · listed by 7 sources
Score
score 703 · 2 ★ · 7 sources · mcp
FAQ (answers come from our records)
How do I install it?
We did not find an install command for this repository (our install coverage is only about 4%). Check its README or releases page.
Who is it for?
We filed it under these tags: MCP 服务. The tags come from the entry itself, not a taxonomy we imposed.
Who maintains it?
repository parastejpal987-cmyk/opticparse-public;first catalogued 2026-10-09;last seen 2026-10-16。There is no official certification or quality rating here — we record sources and dates only.
Third-party quality signals

OpticParse • PhishVision

以下摘录自该仓库的 README(片段) · 在 GitHub 读完整 README →

The Autonomous Agent Web Browsing & Threat Shield Stack
Zero-CSS Visual Extraction • Zero-Day Phishing & Wallet Drainer Defense • MCP Server • Multi-Agent Toolkits
> ⭐ **Support Open Source:** If you find OpticParse or PhishVision useful, please **give us a Star on GitHub**! It helps us maintain free edge scrapers and datasets for everyone.
---
⚡ Feature & Architecture Comparison
Looking for a **Firecrawl alternative**, **Crawl4AI alternative**, or a **Zero-Day Phishing & Visual Shield** for your autonomous AI agents? Here is how OpticParse + PhishVision compares to traditional scraping engines and legacy security blacklists:
| Feature / Capability | OpticParse & PhishVision | Firecrawl | Crawl4AI | Jina Reader | Google Safe Browsing / VirusTotal |
| :--- | :---: | :---: | :---: | :---: | :---: |
| **Primary Extraction Engine** | **Multimodal AI Vision (Zero-CSS)** | HTML / Markdown | DOM Tree / XPath | Regex / Markdown | N/A (Security Only) |
| **Visual Prompt Injection Shield (ToxicCanvas)** | 🛡️ **Built-in (Zero-Opacity & CSS Sanitization)** | ❌ None (Vulnerable) | ❌ None (Vulnerable) | ❌ None | N/A |
| **0-Day Phishing & Wallet Drainer Shield** | 🛡️ **Built-in (Instant Day 0)** | ❌ None | ❌ None | ❌ None | ⚠️ Historical Blacklist (12-48h delay) |
| **Machine Discovery Standard (A2A)** | ✅ **Native agent.json & agent-service.json** | ❌ None | ❌ None | ❌ None | ❌ None |
| **In-IDE Developer Free Trial** | ✅ **200 Requests Out-of-the-Box** | ❌ Mandatory Credit Card | ❌ Self-hosted Setup | ⚠️ Limited Free Tier | ⚠️ Strict Rate Limits |
| **In-Terminal Instant Checkout** | ✅ **ASCII QR in Console + .env Save**| ❌ Web Portal Only | ❌ None | ❌ Web Portal Only | ❌ Enterprise Contract |
| **Agent Ecosystem Toolkits** | **LangChain, ElizaOS, AgentKit, MCP** | LangChain, LlamaIndex | Generic Python | Generic HTTP | None (Raw API) |
| **Payment & Settlement Flexibility** | **USDC (Base/Polygon) / $0.01 M2M** | Stripe ($16+/mo SaaS) | None (Open Source) | Stripe ($$$) | Enterprise ($$$) |
| **Token Optimization Noise Reduction** | **96% Token Reduction** | ~80% | ~70% | ~85% | N/A |
---
📌 Why This Stack Exists: The Dual-Engine Agent Architecture
Giving an autonomous AI agent unrestricted internet access introduces three critical vulnerabilities:
1. **Context Window & DOM Fragility (OpticParse solves this):** Raw HTML consumes 90%+ of LLM token context, while rigid CSS/XPath selectors break the moment websites change frontend frameworks.
2. **Visual Prompt Injection & Screen Poisoning (ToxicCanvas solves this):** Adversarial sites embed invisible CSS, zero-opacity layers, and prompt injection strings to hijack multimodal vision models (Claude Computer Use, GPT-4o, Operator). ToxicCanvas sanitizes DOM tokens before ingestion.
3. **The Autonomous Link Trap (PhishVision solves this):** When agents follow links autonomously, malicious actors exploit them with zero-day credential harvesting kits, fake Web3 dApps, and wallet drainers.
**OpticParse + PhishVision provides the complete solution:**
- **The Shield (PhishVision + ToxicCanvas):** Inspects target URLs in real-time for SSL age, brand spoofing, and visual injection attacks *before* the agent executes navigation.
- **The Engine (OpticParse):** Visually renders authenticated pages at the edge, converting dynamic JavaScript into token-optimized clean Markdown (96% noise reduction) without brittle CSS selectors.
- **Embodied AI & Edge Hardware Compatible:** Lightweight token streams structured for low-power edge processors (NVIDIA Jetson, embedded robotics telemetry) that cannot waste local battery and GPU cycles running headless browsers (Learn more).
Operating **150 autonomous extraction & threat pipelines** across **13 industries**, the network auto-harvests **1,250+ verified intelligence records every 24 hours** at the global edge on Cloudflare Workers, R2, and D1.
Source
awesome · https://github.com/parastejpal987-cmyk/opticparse-public
Browse by field
MCP servers
Traction
2 ★
added
yesterday

More in MCP servers

All 5,525 in MCP servers →

Filed under

Every tag opens the full list for that filter. Counts are the live catalogue figures.

Browse all fields →

← Back to index