There are no articles in the {activeFilter} category yet.
+
+
+ )}
+
+ {/* Newsletter Signup */}
+
+
Subscribe to our newsletter
+
+ Get the latest updates on AutoClaw features, OpenClaw tips, and articles about AI agent deployment.
+
+
+
+
+
+
+ );
+}
diff --git a/website/app/blog/posts.json b/website/app/blog/posts.json
index 0fe17e2..9845ea2 100644
--- a/website/app/blog/posts.json
+++ b/website/app/blog/posts.json
@@ -1,34 +1,38 @@
-{
- "why-openclaw-deployment-hard": {
- "slug": "why-openclaw-deployment-hard",
- "title": "Why Deploying OpenClaw Is Harder Than It Should Be",
- "excerpt": "Exploring the challenges of setting up personal AI assistants: VPS rental, SSH configuration, networking, and why we built AutoClaw.",
- "date": "Jan 28, 2026",
- "readTime": "5 min read",
- "category": "Product"
- },
- "understanding-api-costs": {
- "slug": "understanding-api-costs",
- "title": "Understanding Your AI Agent's API Costs",
- "excerpt": "A breakdown of how API costs accumulate across Anthropic, OpenAI, and other providers—and how to track them effectively.",
- "date": "Jan 15, 2026",
- "readTime": "7 min read",
- "category": "Tutorial"
- },
- "curated-skills-philosophy": {
- "slug": "curated-skills-philosophy",
- "title": "The Case for Curated Skills: Less Is More",
- "excerpt": "OpenClaw has 50+ skills. Most users need 5-10. Here's how we think about skill curation for new users.",
- "date": "Jan 8, 2026",
- "readTime": "4 min read",
- "category": "Philosophy"
- },
- "self-hosted-vs-cloud": {
- "slug": "self-hosted-vs-cloud",
- "title": "Self-Hosted vs Cloud AI: A Decision Framework",
- "excerpt": "Both approaches have trade-offs. This post offers a practical framework for evaluating which makes sense for your situation, with honest pros and cons for each.",
- "date": "Dec 5, 2025",
- "readTime": "5 min read",
- "category": "Product"
- }
-}
+{
+ "why-openclaw-deployment-hard": {
+ "slug": "why-openclaw-deployment-hard",
+ "title": "Why Deploying OpenClaw Is Harder Than It Should Be",
+ "excerpt": "Exploring the challenges of setting up personal AI assistants: VPS rental, SSH configuration, networking, and why we built AutoClaw.",
+ "date": "Jan 28, 2026",
+ "readTime": "5 min read",
+ "category": "Product",
+ "image": "/blog/why-openclaw-deployment-hard.png"
+ },
+ "understanding-api-costs": {
+ "slug": "understanding-api-costs",
+ "title": "Understanding Your AI Agent's API Costs",
+ "excerpt": "A breakdown of how API costs accumulate across Anthropic, OpenAI, and other providers—and how to track them effectively.",
+ "date": "Jan 15, 2026",
+ "readTime": "7 min read",
+ "category": "Tutorial",
+ "image": "/blog/understanding-api-costs.png"
+ },
+ "curated-skills-philosophy": {
+ "slug": "curated-skills-philosophy",
+ "title": "The Case for Curated Skills: Less Is More",
+ "excerpt": "OpenClaw has 50+ skills. Most users need 5-10. Here's how we think about skill curation for new users.",
+ "date": "Jan 8, 2026",
+ "readTime": "4 min read",
+ "category": "Philosophy",
+ "image": "/blog/curated-skills-philosophy.png"
+ },
+ "self-hosted-vs-cloud": {
+ "slug": "self-hosted-vs-cloud",
+ "title": "Self-Hosted vs Cloud AI: A Decision Framework",
+ "excerpt": "Both approaches have trade-offs. This post offers a practical framework for evaluating which makes sense for your situation, with honest pros and cons for each.",
+ "date": "Dec 5, 2025",
+ "readTime": "5 min read",
+ "category": "Product",
+ "image": "/blog/self-hosted-vs-cloud.png"
+ }
+}
diff --git a/website/app/blog/posts/curated-skills-philosophy.md b/website/app/blog/posts/curated-skills-philosophy.md
index 6fd84be..23fd7c7 100644
--- a/website/app/blog/posts/curated-skills-philosophy.md
+++ b/website/app/blog/posts/curated-skills-philosophy.md
@@ -1,35 +1,35 @@
-## The Paradox of Choice
-
-OpenClaw ships with over 50 skills. File management, web browsing, code execution, API integrations, database queries, and much more.
-
-But here's the thing: most users only need 5-10 skills.
-
-## Why More Isn't Better
-
-1. Cognitive overhead - More options means more confusion
-2. Security surface - Each skill is a potential vulnerability
-3. Performance - Unused skills still consume resources
-4. Maintenance - More skills = more things that can break
-
-## Our Curation Philosophy
-
-At AutoClaw, we believe in opinionated defaults:
-
-- Start with essential skills enabled
-- Make it easy to add more when needed
-- Provide clear documentation for each skill
-- Remove skills that cause more problems than they solve
-
-## The Essential Skill Set
-
-For most users, we recommend starting with:
-
-1. File System - Read, write, navigate files
-2. Shell - Execute commands
-3. Web Browser - Research and automation
-4. Code Editor - Write and modify code
-5. Git - Version control
-
-## Conclusion
-
-Don't enable skills "just in case." Start minimal, add as needed. Your AI agent will be faster, safer, and easier to understand.
+## The Paradox of Choice
+
+OpenClaw ships with over 50 skills. File management, web browsing, code execution, API integrations, database queries, and much more.
+
+But here's the thing: most users only need 5-10 skills.
+
+## Why More Isn't Better
+
+1. Cognitive overhead - More options means more confusion
+2. Security surface - Each skill is a potential vulnerability
+3. Performance - Unused skills still consume resources
+4. Maintenance - More skills = more things that can break
+
+## Our Curation Philosophy
+
+At AutoClaw, we believe in opinionated defaults:
+
+- Start with essential skills enabled
+- Make it easy to add more when needed
+- Provide clear documentation for each skill
+- Remove skills that cause more problems than they solve
+
+## The Essential Skill Set
+
+For most users, we recommend starting with:
+
+1. File System - Read, write, navigate files
+2. Shell - Execute commands
+3. Web Browser - Research and automation
+4. Code Editor - Write and modify code
+5. Git - Version control
+
+## Conclusion
+
+Don't enable skills "just in case." Start minimal, add as needed. Your AI agent will be faster, safer, and easier to understand.
diff --git a/website/app/blog/posts/self-hosted-vs-cloud.md b/website/app/blog/posts/self-hosted-vs-cloud.md
index 6027b25..31d2685 100644
--- a/website/app/blog/posts/self-hosted-vs-cloud.md
+++ b/website/app/blog/posts/self-hosted-vs-cloud.md
@@ -1,115 +1,115 @@
-## Evaluating Self-Hosted vs Cloud AI
-
-Both self-hosted and cloud AI can be the right choice. The decision mostly comes down to five variables: privacy requirements, ops comfort, usage volume, time available, and how much customization you need.
-
-This post lays out the real upsides, the real costs, and a simple scoring framework.
-
-## What's Actually Better About Self-Hosting
-
-### 1) Privacy you control (but not absolute)
-
-Self-hosting usually means *you host the app and the data*, but you may still call model APIs unless you run local models.
-
-What self-hosting can meaningfully reduce:
-- Provider visibility into your *full* conversation history (you control retention)
-- Exposure of your files (you choose what gets sent)
-- Workflow/integration context (they see calls, not your whole system)
-
-It matters most for:
-- Proprietary code and business logic
-- Regulated/customer data (GDPR/HIPAA, etc.)
-- Internal tools and databases
-
-It matters less for:
-- Generic questions you'd search anyway
-- Learning/experimentation
-- Public/open-source work
-
-### 2) Cost control (sometimes savings)
-
-Cloud pricing is convenient but can become unpredictable at scale.
-
-Self-hosting can improve cost efficiency via:
-- No reseller markup on API calls
-- Aggressive caching and reuse
-- Right-sized infrastructure
-
-If your usage is sporadic, cloud usually wins.
-
-### 3) Deep customization
-
-Self-hosting lets you tailor the system to your workflow:
-- Custom system prompts and guardrails
-- Custom tools/skills (CI/CD, repos, DB queries)
-- Policies (approvals for sensitive actions, audit logs)
-
-## The Real Costs of Self-Hosting
-
-### 1) Time
-
-Typical costs:
-- Setup: hours to days (depending on experience)
-- Maintenance: ~1–2 hours/month (until something breaks)
-- Security/upgrades: your responsibility
-
-If your time is expensive, this dominates the decision.
-
-### 2) Operational risk
-
-Common failure modes:
-- Outages
-- Backup mistakes
-- Security incidents
-- Dependency/version breakage
-
-Cloud providers absorb most of that.
-
-### 3) Opportunity cost
-
-Every hour on infra is an hour not spent building, learning, serving clients, or resting.
-
-## Decision Framework
-
-Score each 1–5:
-
-- **Privacy requirements** (1: none, 5: strict/compliance)
-- **Technical comfort** (1: avoid servers, 5: production ops experience)
-- **Usage volume** (1: occasional, 5: heavy daily)
-- **Time availability** (1: none, 5: enjoy tinkering)
-- **Customization needs** (1: default is fine, 5: deep integrations)
-
-### Interpretation
-
-- **5–10:** Choose cloud.
-- **11–17:** Choose managed self-hosting (middle ground).
-- **18–25:** Consider full self-hosting (maximum control).
-
-## When Cloud Wins
-
-Cloud is usually better when:
-- You're still learning
-- Your usage is spiky/unpredictable
-- Your time is very expensive
-- You don't want ops work
-- You need enterprise SLAs/support
-
-## When Self-Hosting Wins
-
-Self-hosting is usually better when:
-- You handle sensitive/regulated data
-- AI spend is high and consistent
-- You need custom integrations and policies
-- AI is part of what you sell
-- You prefer ownership and portability
-
-## Questions to Ask Yourself
-
-1. If AI is down for a day, is it a nuisance or a business problem?
-2. What would you do with ~1 hour/week (ongoing maintenance)?
-3. What's actually in your prompts—would exposure matter?
-4. Are you being ideological or solving a concrete requirement?
-5. What's your exit plan (data, workflows, portability)?
-
----
-
-*If you want the benefits of self-hosting without the ops burden, managed self-hosting exists for that middle ground.*
+## Evaluating Self-Hosted vs Cloud AI
+
+Both self-hosted and cloud AI can be the right choice. The decision mostly comes down to five variables: privacy requirements, ops comfort, usage volume, time available, and how much customization you need.
+
+This post lays out the real upsides, the real costs, and a simple scoring framework.
+
+## What's Actually Better About Self-Hosting
+
+### 1) Privacy you control (but not absolute)
+
+Self-hosting usually means *you host the app and the data*, but you may still call model APIs unless you run local models.
+
+What self-hosting can meaningfully reduce:
+- Provider visibility into your *full* conversation history (you control retention)
+- Exposure of your files (you choose what gets sent)
+- Workflow/integration context (they see calls, not your whole system)
+
+It matters most for:
+- Proprietary code and business logic
+- Regulated/customer data (GDPR/HIPAA, etc.)
+- Internal tools and databases
+
+It matters less for:
+- Generic questions you'd search anyway
+- Learning/experimentation
+- Public/open-source work
+
+### 2) Cost control (sometimes savings)
+
+Cloud pricing is convenient but can become unpredictable at scale.
+
+Self-hosting can improve cost efficiency via:
+- No reseller markup on API calls
+- Aggressive caching and reuse
+- Right-sized infrastructure
+
+If your usage is sporadic, cloud usually wins.
+
+### 3) Deep customization
+
+Self-hosting lets you tailor the system to your workflow:
+- Custom system prompts and guardrails
+- Custom tools/skills (CI/CD, repos, DB queries)
+- Policies (approvals for sensitive actions, audit logs)
+
+## The Real Costs of Self-Hosting
+
+### 1) Time
+
+Typical costs:
+- Setup: hours to days (depending on experience)
+- Maintenance: ~1–2 hours/month (until something breaks)
+- Security/upgrades: your responsibility
+
+If your time is expensive, this dominates the decision.
+
+### 2) Operational risk
+
+Common failure modes:
+- Outages
+- Backup mistakes
+- Security incidents
+- Dependency/version breakage
+
+Cloud providers absorb most of that.
+
+### 3) Opportunity cost
+
+Every hour on infra is an hour not spent building, learning, serving clients, or resting.
+
+## Decision Framework
+
+Score each 1–5:
+
+- **Privacy requirements** (1: none, 5: strict/compliance)
+- **Technical comfort** (1: avoid servers, 5: production ops experience)
+- **Usage volume** (1: occasional, 5: heavy daily)
+- **Time availability** (1: none, 5: enjoy tinkering)
+- **Customization needs** (1: default is fine, 5: deep integrations)
+
+### Interpretation
+
+- **5–10:** Choose cloud.
+- **11–17:** Choose managed self-hosting (middle ground).
+- **18–25:** Consider full self-hosting (maximum control).
+
+## When Cloud Wins
+
+Cloud is usually better when:
+- You're still learning
+- Your usage is spiky/unpredictable
+- Your time is very expensive
+- You don't want ops work
+- You need enterprise SLAs/support
+
+## When Self-Hosting Wins
+
+Self-hosting is usually better when:
+- You handle sensitive/regulated data
+- AI spend is high and consistent
+- You need custom integrations and policies
+- AI is part of what you sell
+- You prefer ownership and portability
+
+## Questions to Ask Yourself
+
+1. If AI is down for a day, is it a nuisance or a business problem?
+2. What would you do with ~1 hour/week (ongoing maintenance)?
+3. What's actually in your prompts—would exposure matter?
+4. Are you being ideological or solving a concrete requirement?
+5. What's your exit plan (data, workflows, portability)?
+
+---
+
+*If you want the benefits of self-hosting without the ops burden, managed self-hosting exists for that middle ground.*
diff --git a/website/app/blog/posts/understanding-api-costs.md b/website/app/blog/posts/understanding-api-costs.md
index 63d64e4..0be1a4e 100644
--- a/website/app/blog/posts/understanding-api-costs.md
+++ b/website/app/blog/posts/understanding-api-costs.md
@@ -1,35 +1,35 @@
-## The Hidden Cost of AI Agents
-
-When you run an AI agent, every interaction costs money. Understanding these costs is crucial for budgeting and optimization.
-
-## How Pricing Works
-
-### Anthropic (Claude)
-- Input tokens: $3 per million tokens (Claude 3 Sonnet)
-- Output tokens: $15 per million tokens
-- Context window: 200K tokens
-
-### OpenAI (GPT-4)
-- Input tokens: $10 per million tokens (GPT-4 Turbo)
-- Output tokens: $30 per million tokens
-- Context window: 128K tokens
-
-## Real-World Cost Examples
-
-A typical coding session with an AI agent might involve:
-- 50 back-and-forth messages
-- Average 2,000 input tokens per message
-- Average 1,500 output tokens per response
-
-Total cost: Roughly $2-5 per session depending on the model.
-
-## Tips for Cost Optimization
-
-1. Choose the right model - Use smaller models for simple tasks
-2. Be concise - Shorter prompts = lower costs
-3. Cache when possible - Avoid redundant API calls
-4. Set budgets - Use provider dashboards to set spending limits
-
-## AutoClaw's Cost Tracking
-
-AutoClaw includes built-in cost tracking so you always know exactly what you're spending. No surprises at the end of the month.
+## The Hidden Cost of AI Agents
+
+When you run an AI agent, every interaction costs money. Understanding these costs is crucial for budgeting and optimization.
+
+## How Pricing Works
+
+### Anthropic (Claude)
+- Input tokens: $3 per million tokens (Claude 3 Sonnet)
+- Output tokens: $15 per million tokens
+- Context window: 200K tokens
+
+### OpenAI (GPT-4)
+- Input tokens: $10 per million tokens (GPT-4 Turbo)
+- Output tokens: $30 per million tokens
+- Context window: 128K tokens
+
+## Real-World Cost Examples
+
+A typical coding session with an AI agent might involve:
+- 50 back-and-forth messages
+- Average 2,000 input tokens per message
+- Average 1,500 output tokens per response
+
+Total cost: Roughly $2-5 per session depending on the model.
+
+## Tips for Cost Optimization
+
+1. Choose the right model - Use smaller models for simple tasks
+2. Be concise - Shorter prompts = lower costs
+3. Cache when possible - Avoid redundant API calls
+4. Set budgets - Use provider dashboards to set spending limits
+
+## AutoClaw's Cost Tracking
+
+AutoClaw includes built-in cost tracking so you always know exactly what you're spending. No surprises at the end of the month.
diff --git a/website/app/blog/posts/why-openclaw-deployment-hard.md b/website/app/blog/posts/why-openclaw-deployment-hard.md
index de624a3..f67f715 100644
--- a/website/app/blog/posts/why-openclaw-deployment-hard.md
+++ b/website/app/blog/posts/why-openclaw-deployment-hard.md
@@ -1,28 +1,28 @@
-## The Problem with Self-Hosting AI Agents
-
-Setting up your own AI agent shouldn't require a DevOps degree. Yet, for most users trying to deploy OpenClaw or similar tools, the journey looks something like this:
-
-1. Rent a VPS - Navigate confusing pricing tiers, choose between providers, figure out what specs you actually need
-2. SSH Configuration - Generate keys, manage permissions, deal with firewall rules
-3. Networking - Port forwarding, reverse proxies, SSL certificates
-4. Docker Setup - Install Docker, configure compose files, manage volumes
-5. Maintenance - Updates, backups, monitoring
-
-## Why We Built AutoClaw
-
-AutoClaw eliminates all of this complexity. With a single click, you get:
-
-- Pre-configured VPS with optimal specs for AI workloads
-- Secure networking via Tailscale (no port forwarding needed)
-- Automatic Docker deployment with sensible defaults
-- Built-in monitoring and easy updates
-
-## The Technical Reality
-
-Most users just want their AI agent to work. They don't want to spend hours debugging nginx configurations or figuring out why their WebSocket connections keep dropping.
-
-We've handled the infrastructure so you can focus on what matters: using your AI agent to get things done.
-
-## What's Next
-
-In upcoming posts, we'll dive deeper into specific technical challenges and how AutoClaw solves them. Stay tuned!
+## The Problem with Self-Hosting AI Agents
+
+Setting up your own AI agent shouldn't require a DevOps degree. Yet, for most users trying to deploy OpenClaw or similar tools, the journey looks something like this:
+
+1. Rent a VPS - Navigate confusing pricing tiers, choose between providers, figure out what specs you actually need
+2. SSH Configuration - Generate keys, manage permissions, deal with firewall rules
+3. Networking - Port forwarding, reverse proxies, SSL certificates
+4. Docker Setup - Install Docker, configure compose files, manage volumes
+5. Maintenance - Updates, backups, monitoring
+
+## Why We Built AutoClaw
+
+AutoClaw eliminates all of this complexity. With a single click, you get:
+
+- Pre-configured VPS with optimal specs for AI workloads
+- Secure networking via Tailscale (no port forwarding needed)
+- Automatic Docker deployment with sensible defaults
+- Built-in monitoring and easy updates
+
+## The Technical Reality
+
+Most users just want their AI agent to work. They don't want to spend hours debugging nginx configurations or figuring out why their WebSocket connections keep dropping.
+
+We've handled the infrastructure so you can focus on what matters: using your AI agent to get things done.
+
+## What's Next
+
+In upcoming posts, we'll dive deeper into specific technical challenges and how AutoClaw solves them. Stay tuned!
diff --git a/website/app/components/Sidebar.tsx b/website/app/components/Sidebar.tsx
index c404319..567c1ef 100644
--- a/website/app/components/Sidebar.tsx
+++ b/website/app/components/Sidebar.tsx
@@ -2,12 +2,11 @@
import Link from "next/link";
import Image from "next/image";
import { usePathname } from "next/navigation";
-import { useUser, useClerk } from "@clerk/nextjs";
+import { useUser } from "@clerk/nextjs";
export default function Sidebar() {
const pathname = usePathname();
const { user, isLoaded } = useUser();
- const { signOut } = useClerk();
if (pathname?.startsWith("/call")) {
return null;
@@ -18,7 +17,7 @@ export default function Sidebar() {
// Nav items for signed-out users
const publicNavItems = [
{
- href: "/#hero",
+ href: "/",
icon: (