Ancher
Officially listedAn AI second brain with account-level long-term memory across all saved sources.
Ancher
Ancher is an AI personal knowledge base (second brain) founded by former NewsBreak COO Wu Yongsheng, with former Zhipu AI Chief Strategy Officer Zhang Kuo as chief scientist. It throws everything you read, watch, and save into a semantically recallable memory library, then generates finished outputs grounded in your own material, built for knowledge workers who live on reading and collecting.
Core Capabilities
- Capture everything: Links, PDFs, videos, podcasts, screenshots, voice, and notes in one sweep — no tags, no folders, with browser/Slack/Telegram/email entry points
- Account-level long-term memory: Semantic search and Q&A across all saved content with answers traceable to specific notes — a compounding effect that improves with use
- Generate from your own material: Research briefs, documents, slides, HTML pages, and illustrations; notes can also become podcasts in one click
- Developer interfaces: CLI plus hosted MCP Server plus Agent Skills, compatible with 70+ coding agents including Claude Code/Codex/Cursor
- Multi-platform sync: Web / iOS / Android / Chrome extension, with 10 languages on mobile
- Aggressive pricing: A $0 permanent free tier with all core features; paid plans are only $5/$15 monthly
Use Cases Knowledge workers with heavy information intake who work toward deliverables — researchers, investors, founders, creators, and consultants — especially those tracking fixed topics long-term who need recall across months and sources; also a fit for wiring a knowledge base into AI agent workflows.
Unique Advantages The core difference from NotebookLM is the memory boundary: NotebookLM's memory stops at a single notebook, while Ancher provides account-level long-term memory; that cross-source long-term memory plus agent-callable interfaces is exactly what NotebookLM still lacks.
Editorial Review The founder's track record and chief scientist background cross-check, CLI/MCP/Agent Skills are rare among peers, the differentiation claims are genuinely testable, and pricing is the most aggressive. But English mainstream media coverage is zero, the App Store has only 2 ratings, /security still says Coming soon, and the product has already pivoted once within six months. Recommended for listing, but treat it as early-stage high potential with heavy deflation of claims.
Pricing
### 💰 定价模式：免费增值（Freemium） **起步价**：免费（$0 永久免费层） #### 主要方案 - **Free**：$0 - 永久免费（Always free），含全部核心功能与 100 credits/月，社区支持 - **Digest**：$5/月（年付 $59.99）- 5,000 credits/月 + 邮件支持 - **Pro**：$15/月（年付 $179.99）- 20,000 credits/月 + 优先支持 + 新功能抢先体验 #### 试用/其他信息 无试用期，改为 $0 永久免费层，免费层额度较低（100 credits/月），重生成场景需付费；年付约省 20%。iOS 中国区内购为 Digest ¥38/月（¥398/年）、Pro ¥148/月（¥1,298/年），另有 ¥38/¥68/¥148 三档 credit 加油包；credits 与具体动作的换算规则未公开。⚠️ 目录站流传的 Starter $9.99/Pro $17.99/Premier $24.99 为已废止的 1.0 定价，请以官网 /pricing 为准。 — Visit website
FAQ
Is Ancher free?
There is a real $0 permanent free tier (Always free) with all core features and 100 credits/month; allowances are low and heavy regeneration requires paying, with Digest at $5/month and Pro at $15/month.
What can Ancher be used for?
It is an AI personal knowledge base: capture links, PDFs, videos, podcasts, screenshots, voice, and notes in one place, recall them via semantic search, then generate briefs, documents, slides, HTML pages, and podcasts grounded in your own material.
How does Ancher differ from NotebookLM?
The core difference is the memory boundary: NotebookLM's memory stops at a single notebook, while Ancher provides account-level long-term memory (a compounding effect), plus developer interfaces — CLI/MCP/Agent Skills — that NotebookLM still lacks.
Is Ancher's AI a proprietary model?
No. Ancher is a genuine RAG + LLM application layer; the founder has said a proprietary small model is a future goal. It has not built its own foundation model and does not disclose which underlying models it uses.
Does Ancher support MCP and developer interfaces?
Yes. It offers a CLI (@ancher-ai/cli), a hosted MCP Server, and Agent Skills (one-command npx install), compatible with 70+ coding agents including Claude Code, Codex, and Cursor; the streamify-one GitHub organization verifiably exists.
Does Ancher support Chinese?
The mobile app supports 10 languages including Simplified and Traditional Chinese, but the website and blog are English-first with no Chinese site; Chinese-language user communities have reported missing onboarding and unstable AI output.
Who is Ancher for?
Knowledge workers with very high information intake who work toward deliverables, and heavy AI agent users. Not suitable for those needing team collaboration wikis, mature reputational backing, Chinese-first experiences, or enterprise-grade compliance documentation.
Is Ancher's data secure?
The site promises bank-grade encryption and 'never used to train AI', but the /security and /data-use pages are both Coming soon, with no published encryption standards, data retention periods, or third-party audits — legally, defer to the current privacy policy text.