Search Results

Found results for ""

Sandcastles

Sandcastles

A data-driven research and scriptwriting platform for short-form video creators. Build a Watchlist of channels in your niche, deep-analyze their best-performing videos (1 credit each) to extract transcript, hook, storytelling format, visual layout, and outlier score, then turn those patterns into ideas and scripts. MCP is built in: connect https://mcp.sandcastles.ai/ to Claude Desktop or Google Antigravity, and ask it to bulk-analyze top outliers or research winning hooks and formats directly from the chat. Automations (Pro+) let rules auto-analyze videos by outlier score, views, and engagement. Pricing: a one-time $5/month Starter plan exists for new accounts only, but it is severely limited—10 credits/month, 1 workspace, no MCP, no Automations, no Reports, no new-channel tracking. Full plans: Pro $49/mo (or $39/mo annual, 100 credits, 3 workspaces, MCP), Visionary $99/mo (or $79/mo annual, 250 credits, 5 workspaces), Titan $499/mo (or $399/mo annual, 1,500 credits, 20 workspaces, API access). Note: Starter is a one-time offer and its limitations are easy to miss—full plans start at $49/mo. ScamAdviser shows a trust score of 0 but concludes the site is legit (score likely stale—not scanned in 30+ days); WHOIS is hidden and traffic is low. Not a publishing tool—research and scriptwriting only. Not to be confused with the open-source sandcastle-ai workflow engine on PyPI.

API: 开源: 移动端:
Typefully

Typefully

A social media writing and scheduling tool built around a distraction-free editor, with native support for X (Twitter), LinkedIn, Threads, Bluesky, and Mastodon. It offers thread composition, AI writing suggestions (“Make it punchier”), analytics, and scheduling. Its standout feature is deep AI Agent integration: a full MCP Server (OAuth-based, works with Claude, ChatGPT, Cursor, Notion) and REST API v2 (create, schedule, publish across platforms, upload media, manage tags, webhooks) are available to all users, including the free tier. Free plan: 1 social set, 1 user, 10 posts/month. Paid plans start around $8/mo (third-party data; annual billing roughly $120/year per a user post—verify on official pricing). Note: supports text-first platforms only (no Instagram, TikTok, or YouTube); AI writing features are an add-on to the core writing/scheduling experience, not the primary engine.

API: 开源: 移动端:
GPTZero

GPTZero

One of the earliest and most widely used AI text detectors, founded by Princeton student Edward Tian in January 2023. Detects AI-generated text across models (ChatGPT, Gemini, Claude, Llama, etc.), highlights sentence-level AI content, and offers integrations with Google Docs and a Chrome extension. Current product also includes plagiarism detection, grammar feedback, and writing-process verification. Note: the vendor\'s \"99.9% accuracy\" claim is contested—independent tests have documented false positives on human-written text (e.g., a high-school essay flagged 62% AI; the US Constitution flagged as AI-generated). GPTZero itself states results should not be used to punish or as a final verdict, and that no detector is 100% accurate. Has an ESL de-biasing claim. ARR reported at ~$24M (36Kr, Mar 2026). Free tier available; paid pricing on official site.

API: 开源: 移动端:
MBX Reach

MBX Reach

An email marketing platform by MediaBoxEnt Technologies with a genuinely free plan: 250 contacts, 1,000 emails/month, 1 list, 1 user, 1 automation, no time limit. Paid plans add your own branding, an AI writing assistant (first drafts, subject lines, rewrites), multi-step automations, API/webhooks, an MCP server, and up to 10 independent workspaces for managing multiple brands from one login. Supports sending one campaign in up to six languages (translations share the AI assistant allowance), double opt-in, one-click unsubscribe, and Google Analytics tagging. MBX Reach is an official connector in the Claude directory. Note: AI-generated copy and translations require human review; free plan includes MBX Reach branding in the footer; sending pauses at plan limits until the month changes or you upgrade.

API: 开源: 移动端:
CoinedBrand

CoinedBrand

An AI brand word engineering tool. Enter your industry and product keywords, and it generates brandable names that are likely to be trademarkable and have available domains. Built for founders and creators who want a unique name without hiring an agency. Free to try.

API: 开源: 移动端:
HookSlayer

HookSlayer

A Chrome extension that grows your LinkedIn personal brand with AI. Analyze any profile for a 0-100 power score, generate 5 post drafts in your exact tone, and create AI social visuals—all from a Shadow-DOM panel on the right side of the page. No account needed. Free daily analysis; Premium ($15) and Pro ($29) are one-time payments.

API: 开源: 移动端:
WriteXDM

WriteXDM

Paste any X (Twitter) handle, and WriteXDM reads their real recent posts to write 2 personalized cold DMs in 10 seconds. Built for founders, freelancers, creators, and salespeople who want to start real conversations instead of sending generic pitches. Free options available.

API: 开源: 移动端:
Loudship

Loudship

A monitoring tool for Reddit, Hacker News, and Bluesky. Every morning it emails you the top conversations where someone is asking for a product like yours, each with an AI-generated reply draft based on your product\'s facts. You review and post it yourself to keep your account safe. First digest is free.

API: 开源: 移动端:
Levragd

Levragd

What Levragd does
Levragd is positioned as an AI content department for businesses that need a steady flow of content without assembling a full internal team. The setup starts with business context: your website, product information, customer insights, and brand guidelines. Levragd says it uses that material to build a working model of the business, identify content opportunities, and create drafts.


That positioning matters because the product is aimed at the planning problem as much as the writing problem. A small company may know it needs articles, social posts, or emails, but still lack the time to decide which topics deserve attention, how they should be framed, and how they can be reused.


A practical workflow
A reasonable workflow begins by supplying the company website and brand information. Levragd can then identify opportunities, produce a content output, run a quality check, and support revisions. The source describes checks against brand voice, terminology, and facts, with the user choosing which recommendations to apply before publication.


This makes the product more suitable for supervised content production than for fully automatic publishing. The user remains responsible for reviewing edits, changing drafts, and deciding what goes out. That is a useful operating model for founders and small marketing teams where accuracy and tone still require human judgment.


The listed output types include articles, social posts, and emails. The pricing page also defines a generation broadly: it can be an opportunity batch, a content piece, a quality check, a revision, an idea for another use of existing content, or the Brand DNA model created during setup. Buyers should therefore evaluate the allowance by workflow, rather than assuming every generation means a finished article.


Plans and editorial support
Levragd offers a one-time free allowance of 20 generations with no credit card required. The Desk plan is listed at $99 per month for 75 generations, while Newsroom is listed at $349 per month for 300 generations. Prices are in U.S. dollars and applicable taxes are calculated at checkout according to the supplied pricing material.


There is also a separate Concierge offer at $2,499 per month. It includes the Newsroom allowance plus up to 10 personal editorial reviews per month, voice and tone refinement, clarity and audience review, strategic feedback, a monthly content health check, and a monthly founder review. The offer is described as limited to a small number of clients.


Who should consider it
Levragd appears best suited to solopreneurs, founders, small businesses, and existing content teams that need more structured production capacity. A founder could use it to turn product knowledge into a queue of content opportunities. A small marketing team could use it to draft and check several formats while retaining approval over every publication.


The product is less clearly suited to organizations that need guaranteed autonomous publishing, extensive integrations, or enterprise-grade operational controls. Those capabilities are not established by the supplied material.


Limitations and risks
The main limitation is that AI output still needs review. Levragd explicitly says that human review matters, and its workflow keeps publication under user control. That adds responsibility, but it also reduces the risk of treating generated copy as automatically correct.


Analysis: the generation-based quota may be difficult to forecast for teams that use frequent checks, revisions, and repurposing. Analysis: the quality of the result will depend heavily on the accuracy and completeness of the company information and brand guidance provided during setup. Prospective customers should test representative workflows with their own material before judging whether the plan allowance and editorial depth fit their process.

API: 开源: 移动端:
Homer

Homer

从工作会议到草稿

Homer面向那些希望分享产品背后思考和进展,但又不想反复写下发生了什么的创始人。其核心工作流程是启动会话,选择共享的屏幕、窗口或标签页,然后继续工作。Homer表示,它关注有意义的变化,并利用会话分析寻找可能成为帖子的瞬间。

这使得该工具在工作会议中包含明确决策、变更、教训或障碍时最具相关性:例如,研究后修订产品决策或解决缺陷。这些都是Homer网站上展示的示例,并不保证每次会谈都能产出有用内容。实际价值取决于荷马所指出的时刻是否准确反映了作品本身,以及其草稿是否听起来像将发表这些作品的创始人。

将故事与其目的地相匹配

Homer描述了如何调整故事的结构、节奏和细节程度,适用于X、LinkedIn、Reddit、Threads和构建日志。这些格式的编辑目标不同:X的简洁帖子,LinkedIn的更多背景和教训,以及Reddit的问题和结果的非宣传性描述。构建日志被呈现为更持久的工作记录、学习和未解问题。

一个有用的工作流程是回顾一次会议的瞬间,选择有意义的片段,然后在发布前对草稿进行准确性、隐私和声音的编辑。这种审查很重要:生成的叙述可以将复杂的决定压缩成一个整洁的叙事,而实际的推理或结果可能更不确定。Homer的网站描述的是草稿文章,而非自动发布,因此应将其视为起草辅助工具,而非编辑判断的替代。

适合创始人,但有意权衡

该产品的前提是,创始人通常有值得分享的内容,但可能忘记或难以清晰表达工作中重要的内容。实时记录进展可以帮助人们记录决策和经验,然后将其作为创始人更新或回顾再利用。对于那些在所选屏幕之外进行的工作、未被记录的对话,或没有明确可分享结果的会话,它可能没那么有用。这些是工作流程考虑,而非产品局限。

有意的权衡是观察。Homer 要求用户启动会话并共享选定的屏幕表面。这或许适合专注于编码或设计的会议,但创始人应考虑屏幕是否包含客户信息、凭证信息、私人消息或未发布的内容。该网站表示会话可以暂停、停止或删除,并描述原始帧是私密存储的。这些都是第一方声明;所提供的材料并不能独立说明存储、删除或访问控制在实际中的运作方式。

在依赖它之前需要检查的事项

该网站提供免费课程,作为试用流程的方式,并列出付费套餐,但套餐细节和可用性可能会变化。在做出购买决定前,请查看实时产品页面的最新条款。现有材料未能建立集成、支持的操作系统、发布工作流程或不同工作类型间的性能,因此不要假设。

为了谨慎试验,使用低敏感性会话,并有明显的决策或变化。将检测到的瞬间与实际发生的情况进行比较,纠正任何缺失的上下文,并在核对姓名、数字和声明后才发布。这个测试将证明霍默的捕捉与草稿方法是否节省了足够的回忆和写作精力,从而合理地将其纳入创始人的日常。

API: 开源: 移动端: