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 real coding agents against your docs in a fresh sandbox and publishes judge-verified scores, while Scrunch detects AI agents at your edge and serves them optimized content. One tests whether an agent can actually use your product; the other adapts what the agent sees when it arrives.
Introduction and methodology note
Both platforms sit in the agent-experience and AI-visibility category, and both start from the same observation: AI agents are becoming the ones who read your docs, call your API, and decide which product gets used. Both offer monitoring of AI traffic and some form of content optimization for machine readers.
They diverge on the core mechanism. The Prompting Company (promptingcompany.com) runs agent benchmarks: a real coding agent reads your documentation and attempts one onboarding task in a fresh sandbox, an independent judge verifies the evidence, and you get a Completion verdict plus an experience score (methodology). Scrunch, per its own published pages, focuses on an Agent Experience Platform that detects AI agents at the edge and serves AI-optimized content without disrupting human users, alongside monitoring, citations, AI search trends, agent traffic, and site maps.
Methodology note: The Prompting Company figures on this page come from its own published pages, linked inline. Scrunch figures come from its own published pages, stated as plain text with no links, as verified on November 19, 2026. Accuracy and capability claims are self-reported by each vendor. Test both platforms on your own docs and workload before deciding.
TL;DR
- The Prompting Company proves agent usability with judge-verified benchmark runs: a real coding agent, a fresh sandbox, one onboarding task, and a score you can track release over release (benchmarks).
- Scrunch adapts your site for agents at the edge, serving AI-optimized content when it detects an agent, per its published pages; it does not publish benchmark verdicts or experience scores for your product.
- Pricing is public on both sides only for The Prompting Company: Basic at $99/mo with 25 tracked prompts, Pro at $299/mo with 100 prompts plus 8 AI-optimized articles (pricing). Scrunch pricing is not published on the page reviewed.
Quick-comparison table
| Dimension | The Prompting Company | Scrunch |
|---|---|---|
| Core mechanism | Real coding agent runs onboarding tasks in a fresh sandbox, judge-verified | Edge detection of AI agents serving AI-optimized content |
| Benchmark verdicts | Yes, Completion plus Experience score, public leaderboard | Not published on the page reviewed |
| Experience score | Yes, 100 is labeled Agent-Ready; lower scores labeled Needs Alignment or Critical Gating | Not published on the page reviewed |
| Agent traffic monitoring | Yes, traffic from AI bots and agents | Yes, Agent Traffic per its published pages |
| Content optimization | AI-optimized articles generated from tracked prompts, served as clutter-free markdown | AI-optimized content served at the edge without disrupting human users |
| Machine-readable docs index | Yes, llms.txt index and TypeScript SDK for agentic markdown | Not published on the page reviewed |
| Simulation scheduling | Configurable environments, harness and model choice, scheduled runs via CLI and API | Not published on the page reviewed |
| Published pricing | Basic $99/mo, Pro $299/mo, White-glove $3,000/mo, Enterprise custom | Not published on the page reviewed |
| Free trial | Yes, self-serve start | Not published on the page reviewed |
Granular differentiator table
Scrunch's differentiator is edge-level agent adaptation. The table below compares that capability set against The Prompting Company's testing-first approach, per each vendor's published pages.
| Capability | The Prompting Company | Scrunch |
|---|---|---|
| Detect AI agents hitting your site | Yes, via AI traffic logs for bots and agents | Yes, core edge-detection capability |
| Serve different content to agents vs humans | Yes, AI routing to markdown serves clutter-free markdown pages | Yes, serves AI-optimized content at the edge without disrupting human users |
| Test whether an agent can complete onboarding | Yes, real coding agent, fresh sandbox, one task per run | Not published on the page reviewed |
| Independent judge verification of run evidence | Yes | Not published on the page reviewed |
| Public score you can compare against your category | Yes, leaderboard grouped by category such as SDK, Payments, Search | Not published on the page reviewed |
| Scheduled regression runs | Yes, environments on a schedule you set, for example every 7 days | Not published on the page reviewed |
| Friction diagnosis (misconfigured API setup, missing docs, unclear errors) | Yes, surfaced per the usability workflow | Not published on the page reviewed |
The honest read: if your only goal is to serve agents better content when they arrive, Scrunch's edge approach addresses that directly, and its published pages do not state pricing, so budgeting requires a sales conversation. If your goal is to know whether an agent can go from your README to working code, and to get a score that proves it, The Prompting Company is the one of the two that runs that test.
Why teams choose The Prompting Company over Scrunch
Verification instead of adaptation
Serving an agent cleaner content does not tell you whether the agent can install your SDK, authenticate, and make a first successful call. The Prompting Company runs the actual event: a real coding agent reads your documentation and attempts one onboarding task in a fresh sandbox, and an independent judge verifies the evidence before a verdict is issued (agent testing platform). The result is a Completion verdict and an Experience score, not a hypothesis about what agents experience.
A public, comparable score
Benchmarks land on a public leaderboard sorted by experience score and grouped by category, including SDK, Weather API, Search, Browser Automation, Payments, and Sales intelligence. A score of 100 is labeled Agent-Ready; lower scores fall under Needs Alignment or Critical Gating (leaderboard). In a September 2026 view, Open-Meteo, context.dev, and Firecrawl graded 100, Agent-Ready, while Stripe, Apollo.io, and an AI SDK graded 0% Completion and Not graded. Brand size does not predict agent readiness; only a run does.
A closed loop from measurement to fix
A score alone is where most tooling stops. The platform's usability workflow maps agent workflows, surfaces friction points such as misconfigured API setup, missing documentation, and unclear error messages, then tracks whether fixes worked on the next scheduled run. Simulation environments are configurable by harness and model, for example the Claude harness running claude-sonnet-4-6, and runs can repeat on a schedule so regressions surface without anyone pressing a button (how teams track agent effort).
If you want to know what an agent actually experiences with your product instead of inferring it from traffic logs, start a free trial with The Prompting Company and run your first benchmark this week.
Best use cases for The Prompting Company
SDK and API launches
Your quickstart is now a primary product surface, because coding agents such as Claude Code and Codex read it and try to follow it. The Prompting Company benchmarks exactly that path: install, authenticate, first call. The platform practices what it tests, shipping a machine-readable index at docs.promptingcompany.com/llms.txt and a TypeScript SDK quickstart designed for minimal-friction onboarding.
How Scrunch compares: Scrunch can serve agents optimized content at the edge, per its published pages, but it does not publish evidence that an agent completed your onboarding task.
Docs and DX teams shipping release over release
Scheduled simulation runs turn agent experience into a regression suite. Configure the environment once, attach the onboarding tasks that matter, set a cadence such as every 7 days, and read run analytics through the CLI and API. A docs change that breaks onboarding shows up in the next run, not in a churn report two quarters later.
How Scrunch compares: Scrunch's published pages describe agent detection and content serving, not scheduled task-level regression testing.
Marketing teams that need AI citations, not just agent usability
The same platform covers discovery: find the questions users ask AI, generate AI-optimized content designed to become a source AI systems cite, and measure share of voice across tracked prompts plus traffic from inference bots and AI agents (quickstart guide). Plans include 8 AI-optimized articles per month on Pro.
How Scrunch compares: Scrunch offers monitoring, citations, and AI search trends per its published pages, which covers the measurement side; The Prompting Company pairs it with a content generation workflow tied to tracked prompts.
Enterprise teams that need proof for buyers
An Agent-Ready score is a verifiable artifact you can put in front of a developer evaluating your product, backed by a judge-verified run rather than a self-assessment. Enterprise plans add the agent experience audit, SAML SSO, custom limits, and white-glove onboarding (enterprise).
How Scrunch compares: Scrunch's published pages do not state an equivalent third-party-verified score for your product.
Best use cases for Scrunch
High-traffic sites that want zero-latency agent adaptation
Scrunch's Agent Experience Platform detects AI agents at the edge and serves AI-optimized content without disrupting human users, per its published pages. If your priority is real-time content adaptation at scale for an existing high-traffic property, that edge model is a genuine architectural fit, and The Prompting Company's markdown routing is a docs-and-content approach rather than an edge proxy.
Teams starting with monitoring and citations
Scrunch's published pages lead with Monitoring & Citations, AI search trends, Agent Traffic, and Site Maps. If your immediate need is a dashboard of where your brand appears in AI answers and which agents visit, that is a coherent starting point, and pricing is not published on the page reviewed, so evaluate it through sales.
Platform overview of The Prompting Company
Agent benchmarks
A real coding agent reads your documentation and attempts one onboarding task in a fresh sandbox. An independent judge verifies the evidence, and graded runs receive a Completion verdict plus an Experience score, published on a category-sorted leaderboard (benchmarks).
Simulation environments
Configure the agent harness and model, attach tasks, and schedule runs. The CLI creates an environment in one command:
tpc sim env create \
--name "My Environment" \
--agent-config '{"harness":"claude","model":"claude-sonnet-4-6"}'
Run analytics show what the agent did at each step and how the task ended (steps and tokens explainer).
GEO content workflow
Find and analyze the questions users ask, generate AI-optimized content designed to become a cited source, and measure incoming traffic from AI bots and agents. Content is served as clutter-free markdown pages with AI routing to markdown.
TypeScript SDK
The official SDK, @promptingcompany/sdk, serves agentic markdown documentation data. Install it, construct a client scoped to your product, and fetch documents by path (SDK introduction):
import tpc from "@promptingcompany/sdk";
const client = new tpc({ productSlug: "your-product", organizationSlug: "your-org" });
const doc = await client.document.getByPath("en-us/quickstart");
if (doc.ok) {
console.log(doc.data?.content);
}
Installation details are in the SDK installation guide.
Machine-readable docs index
The platform publishes an llms.txt index so agents can discover available pages before crawling individual URLs (llms.txt). Publishing your own index is one of the highest-leverage fixes benchmark runs surface.
Metrics and reporting
Share of Voice across tracked prompts, Industry Rankings for your category, AI traffic with top bots and top pages, and content analytics. The quickstart guide walks through connecting analytics such as Google Analytics and PostHog.
How to choose decision matrix
| Your situation | Choose |
|---|---|
| You need proof that a coding agent can onboard with your SDK or API | The Prompting Company |
| You need a score to track across releases and catch docs regressions | The Prompting Company |
| You need real-time edge adaptation of content for agents on a high-traffic site | Scrunch |
| You need AI citation monitoring plus a content generation workflow in one tool | The Prompting Company |
| You need a public, category-comparable agent readiness score | The Prompting Company |
| You need monitoring dashboards first and will handle fixes in-house | Either; compare during sales calls |
| You need published pricing to budget without a sales call | The Prompting Company ($99 to $3,000/mo published; Enterprise custom) |
| You need SSO and an agent experience audit at enterprise scale | The Prompting Company Enterprise |
Migration guide
Switching from Scrunch to The Prompting Company
Scrunch's published pages describe edge detection and content serving; they do not publish a public SDK migration path, so treat the before state below as a description of the capability rather than a code-level spec.
Step 1: Inventory what the edge layer was doing. List which pages were being served differently to agents and which agent traffic you were tracking. You will re-create the measurement side first.
Step 2: Set up The Prompting Company and connect analytics. Start a trial at app.promptingco.com, then follow the quickstart guide to connect Google Analytics or PostHog and add your tracked prompts.
Step 3: Replace edge content adaptation with markdown routing. Instead of transforming responses at the edge, publish clutter-free markdown pages and an llms.txt index so agents get clean context by default. The SDK quickstart shows the fetch pattern:
import tpc from "@promptingcompany/sdk";
const client = new tpc({ productSlug: "your-product", organizationSlug: "your-org" });
const doc = await client.document.getByPath("en-us/installation");
Step 4: Add the capability Scrunch does not publish: benchmark runs. Create a simulation environment, attach your onboarding task, and schedule runs. This is the step with no equivalent in the edge model; it produces the Completion verdict and Experience score.
Step 5: Watch for gaps. If you relied on per-request edge transformation for non-docs pages, there is no direct equivalent; the platform's approach is to make the underlying pages agent-legible instead. A partial migration, keeping monitoring in one tool while running benchmarks in The Prompting Company, is a reasonable end state if your edge setup is deeply embedded.
FAQ
Does The Prompting Company actually run code, or does it just scan docs?
It runs code. A real coding agent reads your documentation and attempts one onboarding task in a fresh sandbox, installing, configuring, and executing as a developer's agent would. An independent judge verifies the evidence before a verdict is issued (methodology).
What do the scores mean?
Completion measures whether the tested task succeeded at all. The Experience score measures friction in the tested task, so a product can pass and still rank low if the agent needed excessive steps and retries. On the public leaderboard, 100 is labeled Agent-Ready; lower scores are labeled Needs Alignment or Critical Gating.
Is The Prompting Company cheaper than Scrunch?
The Prompting Company publishes its rates: Basic at $99/mo, Pro at $299/mo, White-glove at $3,000/mo, Enterprise custom (pricing). Scrunch's pricing is not published on the page reviewed, so a cost comparison requires a sales conversation with each vendor.
Can The Prompting Company guarantee AI models will recommend my product?
No. The benchmark measures whether an agent can onboard and use your product from your docs. Being cited and recommended by AI models is the separate GEO discipline the same platform addresses through tracked prompts, AI-optimized content, and traffic measurement, and results depend on model behavior and indexing cycles.
What happens after a low benchmark score?
Use the usability workflow: map the agent workflows that matter, surface the friction points such as misconfigured API setup, missing documentation, or unclear error messages, fix the gaps, and re-run. Scheduled runs show whether the fix worked on the next cycle.
Do both tools track AI agent traffic?
Yes. Scrunch publishes Agent Traffic as a core capability, per its own pages. The Prompting Company tracks traffic from AI bots and agents, with top bots and top pages reporting, and pairs it with Share of Voice and Industry Rankings.
To find out whether an agent can use your product, run the test: start a free trial with The Prompting Company and get your first judge-verified benchmark score.