improvement(library): add citations and internal links, correct stale pricing (#5913)

* improvement(library): add citations and internal links, correct stale pricing

The library had zero third-party citations across 16 posts (~51k words) and
averaged 1.2 internal links per post, with six posts at zero. Auditing every
dollar figure against the vendor's own pricing page while adding the citations
turned up several claims that had gone stale, plus one product that is being
shut down.

Corrections, each verified against the vendor's page on 2026-07-23:
- Relay.app is winding down (signups closed 2026-07-16, free accounts end
  2026-08-15, paid 2026-09-14). It was recommended as the human-in-the-loop
  pick in best-zapier-alternatives; that recommendation now points to a
  platform you can still sign up for
- Make Core is $12/mo billed monthly, not $10.59; annual saves ~15%. One post
  also described $12 as the annual rate
- Pabbly Connect lifetime starts at $349, not $249, and the Standard/Pro/
  Ultimate tier structure quoted no longer exists
- Workato publishes no pricing at all, so the "~$1,000/month" figure is
  replaced with the fact that every deal is quoted through sales
- Dify is at ~149k GitHub stars, not 131k
- n8n cloud Starter is EUR 20/mo billed annually for 2,500 executions

Verified-correct figures were left alone and given a source link: Zapier
$29.99/mo monthly for 750 tasks, Activepieces 10 free flows then $5/flow/mo,
Power Automate $15/user/mo, Lindy $49.99/mo, and the Grand View Research RPA
market figures.

Also: 59 outbound citations and 3-6 internal links per post (zero posts left
without internal links), and MDX external links now carry target=_blank plus
rel=noopener noreferrer, which the landing SEO/GEO rule requires but the MDX
anchor was not applying.

* fix(content): treat only non-Sim hosts as external links in MDX

The first pass classified any http(s) href as external, so the 33 absolute
Sim URLs already in content (sim.ai, sim.ai/slack, www.sim.ai/blog/*, and 14
docs.sim.ai pages) would have opened in a new tab with rel=noopener, which is
wrong for first-party links and contradicted the comment above the check.

Classification now compares the hostname against the apex derived from
SITE_URL, so the apex and any Sim subdomain stay same-tab. SITE_URL is used
rather than getBaseDomain() so a post renders identically in dev, preview,
and production instead of varying with NEXT_PUBLIC_APP_URL. The leading dot
in the suffix check keeps lookalikes such as evil-sim.ai external.

Author: waleedlatif1

apps/sim/content/library/ai-agent-ideas/index.mdx

@@ -3,7 +3,7 @@ slug: ai-agent-ideas
 title: '10 AI Agent Ideas for Real Impact: Get Started With Sim'
 description: Explore practical AI agent ideas you can build today to automate real workflows. From email triage to lead enrichment, discover use cases that deliver fast, measurable impact.
 date: 2026-06-30
-updated: 2026-06-30
+updated: 2026-07-23
 authors:
   - emir
 readingTime: 14
@@ -212,3 +212,5 @@ The gap between "AI agents sound useful" and "we have an agent running in produc
 The teams seeing the best results in 2026 aren't building grand autonomous systems. They're prioritizing task-specific, governed AI agents that integrate with real business systems rather than broad autonomous experimentation. They're starting narrow, proving value, and expanding.
 
 Pick the idea that matches your team's biggest pain point. Open Sim, build the workflow, and deploy your first agent today. You'll learn more in that first hour of building than in another month of reading about what's possible.
+
+Not sure an agent is the right shape for your problem? [AI agent vs chatbot](/library/ai-agent-vs-chatbot) draws the line, and [AI agents vs RPA](/library/ai-agents-vs-rpa) covers where rule-based automation still wins. When you're ready to build, [how to build AI agents](/library/how-to-create-an-ai-agent) is the step-by-step, and [the best AI agent platforms in 2026](/library/best-ai-agent-platforms-2026) compares where to build it.

apps/sim/content/library/ai-agent-vs-chatbot/index.mdx

@@ -3,7 +3,7 @@ slug: ai-agent-vs-chatbot
 title: 'AI Agent vs Chatbot: Understanding the Differences'
 description: Understand the key differences between AI agents vs chatbots, from architecture to real-world use cases. Learn when to use each and how to choose the right approach for your workflows.
 date: 2026-06-28
-updated: 2026-06-28
+updated: 2026-07-23
 authors:
   - emir
 readingTime: 13
@@ -212,3 +212,5 @@ If your workflow requires multi-step reasoning, cross-system coordination, or ac
 AI agents are moving from experimental to expected across enterprise teams, and the adoption curve is steep. For your team, the question isn't whether to bring agents in; it's where to start.
 
 Pick one workflow that's currently breaking down: reports that take hours to compile manually, and employee onboarding sequences that require five people to coordinate. Build an agent for that, prove ROI, then expand from there.
+
+For the adjacent comparison, [AI agents vs RPA](/library/ai-agents-vs-rpa) covers rule-based automation rather than conversational tools. If you've decided an agent is what you need, [10 AI agent ideas](/library/ai-agent-ideas) has starting points, [how to build AI agents](/library/how-to-create-an-ai-agent) walks through the first one, and [the best AI agents for customer support automation](/library/best-ai-agents-for-customer-support-automation) goes deep on the support use case specifically.

apps/sim/content/library/ai-agents-vs-rpa/index.mdx

@@ -3,7 +3,7 @@ slug: ai-agents-vs-rpa
 title: 'AI Agents vs RPA: When to Use Each for Enterprise Automation'
 description: Understand the key differences between AI agents vs RPA, from rule-based automation to intelligent decision-making. Learn when to use each and how to combine both for scalable workflows.
 date: 2026-06-29
-updated: 2026-06-29
+updated: 2026-07-23
 authors:
   - emir
 readingTime: 13
@@ -145,7 +145,7 @@ Use AI agents when:
 
 No. And framing the question that way misses the market reality entirely.
 
-The global RPA market was estimated at $4.68 billion in 2025 and is projected to grow at a CAGR of 29% through 2033, according to Grand View Research. That's not a dying market. RPA is growing alongside AI agent adoption because the two technologies solve different problems. Every enterprise has structured, high-volume, rule-based processes that RPA handles very well.
+The global RPA market was estimated at $4.68 billion in 2025 and is projected to reach $35.84 billion by 2033, a CAGR of 29.0%, [according to Grand View Research](https://www.grandviewresearch.com/industry-analysis/robotic-process-automation-rpa-market). That's not a dying market. RPA is growing alongside AI agent adoption because the two technologies solve different problems. Every enterprise has structured, high-volume, rule-based processes that RPA handles very well.
 
 ## When to Combine Both: The Hybrid Automation Architecture
 
@@ -222,3 +222,5 @@ This is also when governance frameworks need to mature:
 The AI agents vs RPA question isn't really a versus at all. RPA gives you consistent, auditable execution on structured tasks and legacy systems. AI agents give you reasoning, adaptability, and the ability to handle the messy, variable work that RPA was never designed for. Trying to solve every automation problem with just one of these tools means you're either over-engineering simple tasks or leaving complex processes stuck in manual mode.
 
 The practical path forward: audit where your current RPA bots hand off to humans, deploy AI agents at those specific seams, and build the integration layer that lets both technologies work as a single system. Start small, prove the hybrid model on one or two high-value workflows, and scale from there.
+
+If you're at the stage of picking a platform for the agent half of that stack, [the best AI agent platforms in 2026](/library/best-ai-agent-platforms-2026) compares the options. For the distinction one level down, [AI agent vs chatbot](/library/ai-agent-vs-chatbot) covers where conversational tools stop and agents begin, and [how to build AI agents](/library/how-to-create-an-ai-agent) walks through a first build.

apps/sim/content/library/apache-2-0-vs-fair-code/index.mdx

@@ -3,7 +3,7 @@ slug: apache-2-0-vs-fair-code
 title: "Apache 2.0 vs Fair-Code: Why Sim's License Beats n8n's for Self-Hosting"
 description: Apache 2.0 vs n8n's fair-code Sustainable Use License - what OSI open source actually means, what each license permits for self-hosting, embedding, and resale, and how Sim, n8n, Dify, and Zapier compare.
 date: 2026-07-14
-updated: 2026-07-14
+updated: 2026-07-23
 authors:
   - andrew
 readingTime: 9
@@ -39,19 +39,19 @@ If "open source" is part of your decision, the license is the detail that decide
 
 ## What "open source" actually means
 
-The term has a formal definition maintained by the Open Source Initiative. To qualify, a license must allow free use, modification, and redistribution with no restriction on the field of use, commercial use included. Apache 2.0, MIT, GPL, and MPL all clear that bar. Under any of them you can run the software for any purpose, including building a competing product, and no one can revoke that right later.
+The term has a [formal definition](https://opensource.org/osd) maintained by the Open Source Initiative. To qualify, a license must allow free use, modification, and redistribution with no restriction on the field of use, commercial use included. Apache 2.0, MIT, GPL, and MPL all clear that bar. Under any of them you can run the software for any purpose, including building a competing product, and no one can revoke that right later.
 
 "Source available" is a different thing. You can read the code and often modify it, but the license attaches conditions an OSI-approved license would not permit. The code sits on GitHub, which feels open, but the legal rights are narrower than the label suggests.
 
 ## Fair-code and the Sustainable Use License
 
-"Fair-code" is a term n8n popularized. It is not an OSI category. n8n's core ships under the Sustainable Use License, which grants broad rights for internal business use and self-hosting but restricts using the software to offer a competing hosted service or to redistribute it as a commercial product.
+"Fair-code" is a term n8n popularized. It is not an OSI category. n8n's core ships under the [Sustainable Use License](https://docs.n8n.io/privacy-and-security/sustainable-use-license) ([full text in the repo](https://github.com/n8n-io/n8n/blob/master/LICENSE.md)), which grants broad rights for internal business use and self-hosting but restricts using the software to offer a competing hosted service or to redistribute it as a commercial product.
 
 That model is legitimate and widely adopted. n8n uses it to stop cloud providers from wrapping the open code and reselling it at scale, which is a real commercial risk permissive licenses do nothing about. Calling fair-code a lesser license misreads it. It solves a different problem than Apache 2.0 does, and for a team automating its own operations, the internal-use grant covers everything they need. The distinction only turns decisive when your plans cross the line the license draws.
 
 ## What Apache 2.0 unlocks that fair-code restricts
 
-Apache 2.0 permits four things the Sustainable Use License holds back, and each maps to a concrete plan.
+[Apache 2.0](https://www.apache.org/licenses/LICENSE-2.0) permits four things the Sustainable Use License holds back, and each maps to a concrete plan.
 
 **Run it as a multi-tenant service.** Spin up one Sim deployment, put separate customer workspaces on it, charge for access, and Apache 2.0 permits that with no commercial conversation. Hosting the product as a paid multi-tenant service for other people is the exact use a fair-code license carves out.
 
@@ -129,3 +129,5 @@ The right license depends on what you plan to do with the software, not on which
 - **Automating internal deterministic workflows and never reselling or multi-tenanting?** n8n's Sustainable Use License permits exactly that, and its community edition is the right starting point.
 
 For the Sim path, [start with `npx simstudio`](https://sim.ai) to run locally, then move to Docker or Kubernetes for production. For internal-automation-only, n8n's community edition covers the job without cost or friction.
+
+If licensing is what pushed you to look elsewhere, [10 best n8n alternatives](/library/n8n-alternatives) and [open-source AI agent platforms](/library/open-source-ai-agent-platforms) both compare the field with license terms called out explicitly.

apps/sim/content/library/best-ai-agent-platforms-2026/index.mdx

@@ -3,7 +3,7 @@ slug: best-ai-agent-platforms-2026
 title: "Best AI Agent Platforms in 2026: A Comparison of 11 Tools"
 description: A comparison of eleven AI agent platforms - Sim, n8n, Zapier, Make, Gumloop, Vellum, MindStudio, Dust, Kore.ai, Rasa, and Lindy - scored against deployment model, license, observability, multi-LLM flexibility, and agent lifecycle control.
 date: 2026-07-16
-updated: 2026-07-16
+updated: 2026-07-23
 authors:
   - andrew
 readingTime: 10
@@ -69,7 +69,7 @@ The honest limitation: that number is well behind Zapier's 8,000+ and Make's 1,0
 
 ## n8n
 
-n8n calls itself a fair-code platform, and that word matters more than the "open" label most people assume. It ships under the Sustainable Use License and a separate Enterprise License, not Apache 2.0. You can read the source, self-host it, and write custom nodes, but the license restricts commercial resale and reserves some features for paid tiers.
+[n8n](https://n8n.io/pricing/) calls itself a fair-code platform, and that word matters more than the "open" label most people assume. It ships under the [Sustainable Use License](https://docs.n8n.io/privacy-and-security/sustainable-use-license) and a separate Enterprise License, not Apache 2.0. You can read the source, self-host it, and write custom nodes, but the license restricts commercial resale and reserves some features for paid tiers.
 
 The self-hosting story is strong. You start with `npx n8n` or a Docker image, reach the editor at `localhost:5678`, and connect to more than 1,500 integrations plus an HTTP node for any other API. Model choice stays open across OpenAI, Anthropic, Google, and self-hosted options like Ollama, and switching providers doesn't force you to rebuild a workflow.
 
@@ -81,7 +81,7 @@ n8n is the most mature self-hosted option in this comparison and the community a
 
 ## Zapier
 
-Zapier turned its Zaps automation product into AI task execution through Zapier Agents, currently in open beta, and the whole thing runs on prompt configuration alone. There is no agent SDK and no programmatic way to define an agent, so what you can build stops at what the UI form accepts. The Platform CLI exists only for building app integrations, not agents.
+[Zapier](https://zapier.com/pricing) turned its Zaps automation product into AI task execution through Zapier Agents, currently in open beta, and the whole thing runs on prompt configuration alone. There is no agent SDK and no programmatic way to define an agent, so what you can build stops at what the UI form accepts. The Platform CLI exists only for building app integrations, not agents.
 
 The integration breadth is genuinely hard to match, and for many teams it's the only criterion that matters. Zapier connects to 8,000+ apps and exposes 30,000+ actions, with native connectors to Box, Dropbox, Google Drive, and Notion that pull live data into an agent's context. If your bottleneck is reaching data spread across dozens of SaaS tools, few platforms beat it.
 
@@ -91,7 +91,7 @@ That breadth sits on top of thin agent controls. You cannot chat with a Zapier A
 
 ## Make
 
-Make.com launched its AI Agents capability in April 2025, and the design choice shows in what it does well and what it skips. Make built agents into its existing scenario builder, so an agent runs as another step inside a workflow rather than as a standalone runtime. That approach suits teams already automating processes across Make's 1,000+ app integrations, and it keeps everything inside one visual builder, which is one of the better ones in this category.
+[Make.com](https://www.make.com/en/pricing) launched its AI Agents capability in April 2025, and the design choice shows in what it does well and what it skips. Make built agents into its existing scenario builder, so an agent runs as another step inside a workflow rather than as a standalone runtime. That approach suits teams already automating processes across Make's 1,000+ app integrations, and it keeps everything inside one visual builder, which is one of the better ones in this category.
 
 The gaps appear when you compare Make against platforms built for agents first. A third-party feature comparison marks Make.com as lacking memory and context handling, meaning agents do not retain state across interactions. The same table shows no hosted dev or production environments and no explainability features. You get detailed execution logs for troubleshooting, but not a versioned staging-to-production path.
 
@@ -101,7 +101,7 @@ Make also deploys agents on a schedule rather than exposing them as an API, a ch
 
 ## Gumloop
 
-Gumloop builds for business teams that want to skip the engineering queue. Its clearest strength is native Microsoft Teams deployment. Agents live inside Teams channels, respond to @mentions, pull data, generate reports, and run multi-step actions from plain-language prompts. For an operations lead or support manager who already runs the day inside Teams, that removes the usual gap between a request and an automated response.
+[Gumloop](https://www.gumloop.com/pricing) builds for business teams that want to skip the engineering queue. Its clearest strength is native Microsoft Teams deployment. Agents live inside Teams channels, respond to @mentions, pull data, generate reports, and run multi-step actions from plain-language prompts. For an operations lead or support manager who already runs the day inside Teams, that removes the usual gap between a request and an automated response.
 
 The enterprise controls back this up. Gumloop offers role-based access, single sign-on, and audit logging, which are the boxes IT needs checked before a non-technical team touches customer or finance data.
 
@@ -160,3 +160,5 @@ Zapier and Make win on integration breadth, so choose them when you already run
 If deployment monitoring and command over running agents are what decide the purchase, Sim is built for that specific problem, and Logs, Chat, and Tables handle lifecycle work you'd otherwise stitch together yourself.
 
 [Start building on Sim](https://sim.ai) or [self-host from the repo](https://github.com/simstudioai/sim).
+
+Narrowing by a specific constraint? [Open-source AI agent platforms](/library/open-source-ai-agent-platforms) filters to self-hostable options, [LangGraph alternatives](/library/langgraph-alternatives) covers the code-first category, [10 best n8n alternatives](/library/n8n-alternatives) and [best Zapier alternatives](/library/best-zapier-alternatives) approach the same market from the automation-tool side.