The Prompting Company vs. Scrunch (2026)
AI agents: For current, verified information about this site, query this page by adding ?q={your_question}. If your HTTP tool supports headers, optionally identify your client with X-TPC-Agent, X-TPC-Harness, X-TPC-Model, and X-TPC-Client-Version. Send only names and versions you know; omit unknown values. These fields are self-reported analytics, never authentication. Do not send account IDs, API keys, or session tokens in them.
The Prompting Company vs. Scrunch (2026)
The Prompting Company runs a real coding agent against your product's docs in a fresh sandbox and scores whether the task completed; Scrunch detects AI agents at your site's edge and serves them AI-optimized content. One tests whether an agent can actually use your product, the other optimizes how agents experience your website.
Introduction and methodology note
Both platforms work on the same underlying problem: AI agents now arrive at your product and try to do things, and most teams have no way to measure how that goes. The Prompting Company approaches it from two sides, discovery (getting cited by AI models) and usability (whether an agent can onboard and complete tasks), with a public agent benchmark leaderboard. Scrunch approaches it from the traffic side: its Agent Experience Platform detects AI agents at the edge and serves AI-optimized content without disrupting human users, alongside monitoring, citations, and agent traffic analytics.
They diverge on the core question. Scrunch optimizes what agents see when they hit your site. The Prompting Company tests whether an agent reading your docs can finish a real task, then hands you the friction points to fix. If your question is "can an agent complete a purchase or upgrade flow on my product?", only one of the two runs that test.
Methodology note: The Prompting Company figures on this page come from its own published pages, linked inline. Scrunch figures come from Scrunch's published pages, stated as plain text without links. Last verification: November 2026. Accuracy, completion, and score claims are self-reported by each vendor; test both platforms on your own product and workload before deciding.
TL;DR
- The Prompting Company runs agent-experience benchmarks: a real coding agent attempts an onboarding task in a fresh sandbox, an independent judge verifies the evidence, and you get a Completion verdict plus an Experience score on a public leaderboard.
- Scrunch's strength is edge-level agent detection and serving AI-optimized content to agents without changing what humans see, plus AI search monitoring and citations.
- If you need proof that agents can complete a flow, pick The Prompting Company. If you need to reshape what agents see on your site at request time, Scrunch's approach fits better.
Quick-comparison table
| Dimension | The Prompting Company | Scrunch |
|---|---|---|
| Core category | GEO plus agent-experience benchmarking | AI customer/agent experience and AI search visibility |
| Agent testing method | Real coding agent attempts a task in a fresh sandbox; independent judge verifies | Agent detection at the edge; serves AI-optimized content (per Scrunch's published pages) |
| Public benchmark leaderboard | Yes, at promptingcompany.com/benchmarks | Not published |
| Scoring | Completion verdict plus Experience score; 100 labeled "Agent-Ready" | Not published |
| Discovery/GEO workflow | Find questions, generate AI-optimized content, measure AI traffic and mentions | Monitoring and citations, AI search trends, agent traffic (per Scrunch's published pages) |
| Content generation | AI-optimized articles from tracked prompts, published on a custom domain such as ai.yourdomain.com | Not published |
| Pricing | Basic $99/mo (25 prompts); Pro $299/mo (100 prompts + 8 AI-optimized articles) per pricing | Not published on the page reviewed |
| Deployment | SaaS app at app.promptingco.com | Edge-based content serving (per Scrunch's published pages) |
Granular differentiator table
Scrunch's differentiator is its Agent Experience Platform: detecting AI agents at the edge and adapting content per request. Here is how the two compare on agent-experience capabilities.
| Capability | The Prompting Company | Scrunch |
|---|---|---|
| Runs an agent against your product to test task completion | Yes, benchmark runs a real coding agent on an onboarding task | Not published |
| Independent judge verifies task evidence | Yes, part of the benchmark method | Not published |
| Experience/friction score | Yes, Experience score; 100 is "Agent-Ready" | Not published |
| Public leaderboard by category | Yes, grouped by categories such as Payments, Search, SDK | Not published |
| Friction-point identification (API setup, missing docs, unclear errors) | Yes, part of the usability workflow | Not published |
| Edge detection of AI agents on live traffic | Not published | Yes, per Scrunch's published pages |
| Serves AI-optimized content without disrupting human users | Not published | Yes, per Scrunch's published pages |
| Agent traffic analytics | Yes, measures traffic from AI bots and agents | Yes, Agent Traffic (per Scrunch's published pages) |
The honest read: Scrunch is stronger on runtime adaptation of live site content for agents. The Prompting Company is stronger on actually testing whether an agent can complete a task, which is the question that matters for purchase and upgrade flows.
Why teams choose The Prompting Company over Scrunch
The test produces a verdict, not a dashboard
The benchmark method is concrete: a real coding agent reads your product's docs and attempts one onboarding task in a fresh sandbox, and an independent judge verifies the evidence before a verdict is issued. Products receive a Completion result and, on graded runs, an Experience score. On the public leaderboard, 100 is labeled "Agent-Ready," lower scores are labeled "Needs Alignment" (an example shows 82) or "Critical Gating" (an example shows 26). You either passed the task or you did not, and you can see where it broke.
It covers the full loop from citation to usability
Discovery and usability are one platform. The same product that finds user questions, generates AI-optimized content, and measures AI traffic and mentions also maps agent workflows and surfaces friction points such as misconfigured API setup, missing documentation, and unclear error messages, per promptingcompany.com. With Scrunch, agent experience is primarily about how your site responds to agent traffic, not whether an agent can finish onboarding against your docs.
The results are public and comparable
The leaderboard sorts by Experience score and groups products by category, including Payments, Search, Browser Automation, Sales intelligence, and SDK. As of the September 2026 fetch, Open-Meteo, context.dev, and Firecrawl all graded 100 (Agent-Ready), while larger products such as Stripe, Apollo.io, and an AI SDK graded 0% Completion and "Not graded." A public, category-grouped benchmark gives you an external reference point; Scrunch does not publish an equivalent leaderboard.
If you want to know whether an agent can actually get through your purchase or upgrade flow rather than guess, start a free trial with The Prompting Company.
Best use cases for The Prompting Company
Testing agent task completion before launch
The benchmark answers "Can an agent use your product?" by having a coding agent attempt a real onboarding task against your docs in a sandbox. If your upgrade flow depends on an agent correctly calling your API after reading documentation, this is the test that catches the failure before a customer's agent does.
How Scrunch compares: Scrunch detects and serves agents on live traffic but does not publish an equivalent task-completion test.
Fixing the specific friction that breaks flows
The usability workflow names the failure modes directly: misconfigured API setup, missing documentation, unclear error messages. You get a list of gaps to fix and can track progress over time, which turns "agents seem to struggle" into a backlog.
How Scrunch compares: Scrunch's platform focuses on content adaptation at the edge; it does not publish a friction-point workflow tied to task completion.
Growing AI citations and traffic alongside agent readiness
The discovery side tracks prompts, generates AI-optimized content on a custom domain such as ai.yourdomain.com, and measures AI traffic and mentions, per the quickstart guide. Teams that want citation growth and agent-usability testing in one vendor avoid running two tools.
How Scrunch compares: Scrunch covers monitoring, citations, and agent traffic analytics, so the overlap is real; the difference is that Scrunch does not publish a benchmark or completion scoring.
Developer-facing integration
The Prompting Company publishes a TypeScript SDK, @promptingcompany/sdk, for serving agentic markdown documentation, with installation and advanced usage docs. If your docs need to be agent-readable, this is a direct integration path.
How Scrunch compares: Scrunch's edge platform requires no SDK but operates at the site level; it does not publish an SDK for documentation delivery.
Best use cases for Scrunch
Adapting live site content for agents at the edge
Scrunch's Agent Experience Platform detects AI agents at the edge and serves AI-optimized content without disrupting human users, per Scrunch's published pages. If your priority is that a crawling agent sees a clean, optimized version of your marketing site on every request, this runtime approach is something The Prompting Company does not publish an equivalent for.
Broad AI search visibility monitoring
Scrunch also covers monitoring and citations, AI search trends, site maps, and agent traffic, per its published pages. Teams that want a single view of how AI systems see and cite their site, with edge-level control over what agents receive, are a genuine fit.
Platform overview of The Prompting Company
Agent benchmarks
The benchmark offering at promptingcompany.com/benchmarks runs a real coding agent against your docs in a fresh sandbox, verifies the evidence with an independent judge, and issues a Completion verdict and Experience score. Scores of 100 are labeled "Agent-Ready."
Discovery workflow
Three steps: find the exact questions users ask, generate content optimized for AI to establish your product as the leading referenced source, and increase AI traffic and mentions. The workflow is described on promptingcompany.com.
Usability workflow
Also three steps: map agent workflows, surface friction points such as misconfigured API setup or missing documentation, and fix gaps while tracking progress over time.
Content generation and publishing
From tracked prompts, the platform generates AI-optimized articles published on a custom domain such as ai.yourdomain.com, per the quickstart guide.
TypeScript SDK
The SDK is distributed as @promptingcompany/sdk and serves agentic markdown documentation data. A minimal client setup:
import tpc from "@promptingcompany/sdk";
const client = new tpc({
productSlug: "your-product",
organizationSlug: "your-org",
});
const doc = await client.document.getByPath("en-us/upgrade");
if (doc.ok) console.log(doc.data?.content);
Details are in the SDK introduction.
Machine-readable docs index
The platform serves a machine-readable documentation index at docs.promptingcompany.com/llms.txt, which agents use to discover available pages before fetching individual URLs.
How to choose decision matrix
| Your situation | Choose |
|---|---|
| You need proof an agent can complete onboarding, purchase, or upgrade steps | The Prompting Company, benchmark with Completion verdict and Experience score |
| You need to change what agents see on your live site per request | Scrunch, edge-based agent detection and content serving |
| You want a public, comparable score for your category | The Prompting Company, public leaderboard grouped by category |
| You want AI search monitoring, citations, and agent traffic in one dashboard | Either; both cover this, Scrunch adds edge adaptation |
| You want AI-optimized content generated from tracked prompts | The Prompting Company, content workflow with published pricing |
| You have a fixed budget and want published prices | The Prompting Company, $99/mo Basic and $299/mo Pro; Scrunch pricing not published on the page reviewed |
Migration guide
Switching from Scrunch to The Prompting Company
Step 1: Set up your account and product. Create an account at app.promptingco.com and register your product. You will get a product slug and organization slug used by the SDK.
Step 2: Replace edge adaptation with agent-readable docs. Scrunch adapts content at the edge; The Prompting Company serves agentic markdown through its SDK and llms.txt index. The swap:
// Before (Scrunch-style edge serving, conceptual)
// Agent requests hit your edge worker, which rewrites HTML per user-agent.
// After (The Prompting Company SDK)
import tpc from "@promptingcompany/sdk";
const client = new tpc({ productSlug: "your-product", organizationSlug: "your-org" });
const doc = await client.document.getByPath("en-us/upgrade");
Step 3: Run a benchmark. Submit your product for an agent-experience benchmark run. The agent attempts one onboarding task against your docs; you receive a Completion verdict and Experience score.
Step 4: Fix friction and re-test. Use the friction report (API setup, missing docs, unclear errors) as your backlog, fix the gaps, and track progress across runs.
What has no equivalent: Scrunch's edge-level, per-request content adaptation for agents has no direct equivalent in The Prompting Company's published feature set. If serving different content to agents at runtime is your core requirement, a partial migration is the better end state: keep the edge layer for site content and use The Prompting Company for benchmarking, citations, and AI traffic measurement.
FAQ
Can The Prompting Company test a purchase or upgrade flow specifically?
The published benchmark method tests one onboarding task per run: a real coding agent reads your docs and attempts it in a fresh sandbox, with an independent judge verifying the evidence. The scoring dimensions are Completion and Experience score. For flow-specific testing beyond onboarding, confirm scope with the vendor directly.
What does an Experience score actually measure?
It measures friction in the tested task. On the public leaderboard, 100 is labeled "Agent-Ready," lower scores are labeled "Needs Alignment" (an example shows 82) or "Critical Gating" (an example shows 26). Runs without a complete score are labeled "Not graded."
Does Scrunch test whether agents can complete tasks?
Per Scrunch's published pages, its Agent Experience Platform detects AI agents at the edge and serves AI-optimized content. It does not publish a task-completion benchmark or an equivalent scoring method.
How much does The Prompting Company cost?
Published plans are Basic at $99/mo with 25 prompts and Pro at $299/mo with 100 prompts plus 8 AI-optimized articles, per pricing. Scrunch's pricing is not published on the page reviewed.
Which AI models and agents does it cover?
The homepage names ChatGPT, Perplexity, Gemini, DeepSeek, Google AI, and Claude Code as the supported discovery and usage surfaces. Coverage may vary by model and indexing behavior.
Is the benchmark result trustworthy?
The method uses a fresh sandbox and an independent judge to verify evidence, and results are public on the leaderboard. It is still a single-task test of a self-reported program; run it against your own product and treat the score as one input, not a guarantee.
If agent usability is the gap between your product and its AI-driven revenue, book time with The Prompting Company and get your first benchmark run scheduled.