The Prompting Company vs. Profound (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. Profound (2026)
The Prompting Company proves whether coding agents can actually use your MCP server with task-based benchmark runs, while Profound's published feature set stops at observing AI crawler traffic.
Introduction and methodology note
An MCP server gives coding agents a way to call your product. It does not tell you whether any agent ever found it, tried it, or gave up halfway through onboarding. Those are three different outcomes, and standard analytics cannot separate them because they are built for human sessions and usually filter bot traffic out entirely.
The Prompting Company (promptingcompany.com) is an agent experience platform built for devtools and software products with an MCP server, CLI, API, SDK, or docs-based workflow. It tracks AI agent and inference bot traffic in real time (quickstart guide), serves agents clutter-free markdown pages, and runs agent benchmarks that grade whether a coding agent can complete onboarding with your product. Profound is an enterprise AI visibility platform. Per its own published feature pages, it combines prompt volumes, answer engine insights for brand mentions and sentiment, Agent Analytics for AI crawler activity and traffic, and content workflow agents. The overlap is AI traffic measurement. The divergence is that The Prompting Company tests whether an agent can actually use your product, while Profound's published feature set stops at observing crawler activity.
Methodology note: all The Prompting Company figures on this page come from its own published pages, linked inline, last verified November 2026. Profound figures come from its own published feature pages, stated as plain text without links. Accuracy and coverage claims are self-reported by each vendor. Test both platforms on your own docs, MCP server, and workload before deciding.
TL;DR
- Discovery and usage are separate signals. Server logs and crawler analytics show which AI agents reach your endpoints; only a task-based test shows whether an agent can finish onboarding with your MCP server.
- The Prompting Company covers both: real-time AI traffic analytics with top bots and request-level logs, plus agent benchmarks that run a real coding agent against your docs in a fresh sandbox and return a Completion result and an Experience score.
- Profound is the stronger pick if your primary need is enterprise prompt-volume and brand-mention intelligence across many engines. If your question is "can a coding agent actually use my MCP server," The Prompting Company is the one that answers it.
Quick-comparison table
| Dimension | The Prompting Company | Profound |
|---|---|---|
| Core category | Agent experience platform (discovery + usability) | Enterprise AI visibility platform |
| Task-based agent benchmark | Yes, real coding agent, fresh sandbox, independent judge (benchmarks) | Not published |
| Agent experience score | Yes, Completion result plus Experience score, 100 labeled Agent-Ready (benchmarks) | Not published |
| AI crawler / agent traffic analytics | Yes, real-time hits from AI agents, crawlers, and search bots, with top bots, top pages, and request-level logs (quickstart guide) | Yes, per Profound's published feature pages, Agent Analytics covers AI crawler activity and traffic |
| Prompt volume data | Tracked prompts with suggested prompts from your product and traffic (quickstart guide) | Yes, per Profound's published pages, prompt volumes based on 1.3B+ real user conversations |
| Share of voice and rankings | Yes, Share of Voice, Industry Rankings, proprietary Visibility Score (competitor analysis) | Yes, per Profound's published pages, brand mentions, sentiment, competitor benchmarking |
| Markdown delivery for agents | Yes, AI routing to markdown plus a machine-readable llms.txt index | Not published |
| Simulation scheduling | Yes, configurable agent harness and model, runs on a schedule via CLI (how teams track agent effort) | Not published |
| Coding agents tested | Claude Code, Codex, OpenCode (MCP server FAQ) | Not published |
| Discovery surfaces | ChatGPT, Gemini, Perplexity, Claude, Google AI, DeepSeek (MCP server FAQ) | Broader engine list per Profound's published pages, including ChatGPT, Gemini, Claude, Perplexity, and AI Overviews |
| Published pricing | Basic $99/mo with 25 tracked prompts; Pro $299/mo with 100 prompts plus 8 AI-optimized articles (pricing) | Not published on the pages reviewed |
| Free trial | Yes, start free trial | Not published on the pages reviewed |
Granular differentiator table
Profound's differentiator is enterprise visibility intelligence at scale. The differentiator The Prompting Company owns is proving usability, not just observing traffic. This table compares the two on the agent-experience dimension specifically.
| Capability | The Prompting Company | Profound |
|---|---|---|
| Real coding agent attempts onboarding | Yes, agent reads your docs and attempts one task in a fresh sandbox | Not published |
| Independent judge verifies evidence | Yes, verdict published only after judge verification | Not published |
| Completion result | Yes, whether the tested task succeeded | Not published |
| Experience score | Yes, measures friction in the tested task; 100 is Agent-Ready, lower scores are Needs Alignment or Critical Gating | Not published |
| Public leaderboard by category | Yes, grouped by categories such as SDK, payments, search, and browser automation | Not published |
| Configurable simulation environments | Yes, CLI accepts agent harness and model, for example the Claude harness running claude-sonnet-4-6 | Not published |
| Scheduled retesting | Yes, environments run on a schedule so regressions surface after each docs change | Not published |
| Friction surfacing | Yes, flags misconfigured API setup, missing documentation, and unclear error messages | Not published |
| AI crawler traffic observation | Yes, request-level logs of AI agents, crawlers, and search bots | Yes, Agent Analytics per Profound's published feature pages |
On raw crawler-traffic observation the two platforms overlap. On everything above that line, Profound's published feature pages describe no equivalent, and The Prompting Company's benchmarks are the only published, task-based answer to whether an agent can use your MCP server.
Why teams choose The Prompting Company over Profound
Crawler traffic is a leading indicator, not a verdict
Profound's Agent Analytics tells you that AI crawlers hit your site. That answers the discovery question. It cannot tell you that an agent fetched your MCP docs, misread an auth step, retried three times, and gave up. The Prompting Company's benchmarks run that exact scenario: 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 published. Discovery traffic without completed tasks is a warning, not a win, and only a task-based test distinguishes the two.
The score is built for prioritization, not reporting
Graded runs produce two numbers. Completion measures whether the tested task succeeded. Experience measures friction, which is what extra steps and burned tokens look like from the outside. On the public leaderboard, 100 is labeled Agent-Ready, lower scores fall under Needs Alignment or Critical Gating, and runs without a complete score are Not graded. That gives your team a ranked fix list instead of a dashboard. The usability workflow maps agent workflows, surfaces friction points such as misconfigured API setup, missing documentation, and unclear error messages, then tracks progress as you fix them.
Fixes are verifiable on a schedule
Simulation environments are configurable from the CLI, including the agent harness and model, and run on a schedule you set. Attach your onboarding task IDs, set a weekly run, and every docs change gets retested without anyone pressing a button. Profound's published feature pages describe content workflow agents, but no published equivalent of scheduled, graded agent runs against your own product. If your MCP server is meant to be wired up by agents, start a free trial and get a baseline score this week.
Best use cases for The Prompting Company
You just launched an MCP server and need a baseline
The benchmark answers the launch question directly: can an agent discover, understand, and complete a task with your server using only your public docs? You get a verdict, an Experience score, and the specific friction that blocked the run. How Profound compares: it can show whether AI crawlers visited your docs, but per its published feature pages it offers no task-based test of whether onboarding succeeds.
You ship docs and SDKs that agents read
The platform serves clutter-free markdown pages through AI routing to markdown, publishes a machine-readable llms.txt index so agents can discover every page, and exposes the same agentic markdown data through a TypeScript SDK. How Profound compares: its published pages describe content workflows for your team, not agent-facing delivery of your docs.
You need to know whether AI models recommend you at all
Tracked prompts run across ChatGPT, Gemini, Perplexity, and Claude produce Share of Voice, Industry Rankings, and a proprietary Visibility Score, so you can see the prompts where a rival gets cited instead of you. The competitor analysis view ties every gap to the exact prompt where it shows up. How Profound compares: this is Profound's home turf, and its prompt volumes from 1.3B+ real user conversations are a genuine strength for demand-side question research.
You want regression detection, not a one-time audit
Scheduled simulation runs retest your onboarding flow after every docs change, so a fix that ships shows up in the next run and a regression surfaces without anyone noticing manually. How Profound compares: no published equivalent of scheduled, graded agent runs.
Best use cases for Profound
Enterprise brand-visibility intelligence at scale
Per Profound's published pages, its prompt volumes are based on 1.3B+ real user conversations, and Answer Engine Insights covers brand mentions, sentiment, and competitor benchmarking. For an enterprise marketing team whose primary question is demand-side prompt research and brand perception across many answer engines, that dataset is a real advantage The Prompting Company does not claim.
Broad multi-engine monitoring programs
Profound's published pages describe coverage across a broader list of engines, including AI Overviews alongside ChatGPT, Gemini, Claude, and Perplexity. If your mandate is a wide net of engines and executive reporting rather than agent usability, Profound fits that brief. Its published feature pages do not describe testing whether a coding agent can complete a task with your product.
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 result plus an Experience score. The public leaderboard groups products by category and labels 100 as Agent-Ready.
AI traffic analytics
The platform measures incoming traffic from AI agents, crawlers, and search bots in real time, with total visits, traffic graphs by model, top bots, top pages, and request-level logs. High traffic from inference bots such as OpenAI User confirms your content is being served inside live AI chats. The quickstart guide walks through reading these reports.
AI-optimized delivery
AI routing to markdown serves agents clutter-free markdown pages instead of navigation menus and scripts, and the llms.txt index makes every page discoverable. No product changes are required to start.
TypeScript SDK
The official SDK, @promptingcompany/sdk, delivers agentic markdown data to your own tooling. Install it, construct a client with your product and organization slugs, and fetch documents by path:
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);
Setup details are in the SDK quickstart.
Discovery measurement
Tracked prompts, Share of Voice, Industry Rankings, and a proprietary Visibility Score show whether AI models mention your product across ChatGPT, Gemini, Perplexity, and Claude. The competitor analysis dashboard shows the prompts where rivals win.
Simulation environments
Configure the agent harness and model, attach tasks, and schedule runs through the CLI:
tpc sim env create \
--name "MCP onboarding" \
--agent-config '{"harness":"claude","model":"claude-sonnet-4-6"}'
Run analytics show what the agent did at each step and how the task ended, so a failed run is a debugging artifact, not a mystery.
How to choose decision matrix
| Your situation | Choose |
|---|---|
| You launched an MCP server and need to know if agents can use it | The Prompting Company, for the task-based benchmark and Experience score |
| You need enterprise prompt-volume and sentiment research across many engines | Profound, for its 1.3B+ conversation prompt volumes and Answer Engine Insights |
| You want regression detection on agent onboarding after every docs change | The Prompting Company, for scheduled simulation runs |
| Your primary buyer question is brand visibility in AI answers, not agent usability | Profound |
| You want agent-facing delivery: markdown routing, llms.txt, SDK | The Prompting Company |
| You need a public, judge-verified score you can point customers at | The Prompting Company, via the benchmarks leaderboard |
| Budget matters and you want published self-serve pricing | The Prompting Company, Basic at $99/mo and Pro at $299/mo; Profound's pricing is not published on the pages reviewed |
Migration guide: switching from Profound to The Prompting Company
Most teams do not rip anything out. Profound's visibility reporting and The Prompting Company's agent-experience testing answer different questions, and a partial migration that keeps both is a legitimate end state. If you are consolidating, here is the path.
- Create your account. Start a free trial and connect your product. No code changes to your product are required.
- Rebuild your prompt set. Recreate your highest-value tracked prompts in the Prompts tab. The platform suggests 10 prompts per day based on your product and traffic, and you accept or reject them so the suggestions improve.
- Point agent delivery at the platform. Enable AI routing to markdown and publish your llms.txt index so agents get clean pages. If you served agents through Profound's site-preparation features, verify each page renders as markdown after the switch.
- Run your first benchmark. Attach your MCP onboarding task and read the Completion result and Experience score. This is the step with no Profound equivalent, and it is usually where the surprises are.
- Schedule retesting. Create a simulation environment with your chosen harness and model:
tpc sim env create \
--name "Weekly MCP check" \
--agent-config '{"harness":"claude","model":"claude-sonnet-4-6"}'
What to watch for: Profound's prompt-volume dataset has no equivalent here, so if demand-side question research is central to your workflow, keep Profound for that and use The Prompting Company for benchmarks, AI traffic, and delivery. Friction fixes are on you either way. The platform shows the gaps; you own the fixes.
FAQ
Can't I just check my server logs to see whether agents use my MCP server?
Logs identify which AI agents and crawlers reach your endpoints, which answers discovery. They cannot show where an agent stalled, what confused it, or why it gave up, and they say nothing about agents that never tried your product at all. The benchmarks answer the usage question logs cannot.
What exactly does the agent experience score measure?
Graded benchmark runs produce a Completion result and an Experience score. Completion measures whether the tested task succeeded. Experience measures friction in the tested task. A score of 100 on the public leaderboard is labeled Agent-Ready, and lower scores are labeled Needs Alignment or Critical Gating.
Which coding agents do the usability tests run against?
Usability testing covers Claude Code, Codex, and OpenCode. Discovery measurement spans ChatGPT, Claude, Perplexity, Google AI, Gemini, and DeepSeek, so you can see both whether agents find you and whether they can use you.
Is Profound cheaper than The Prompting Company?
Not published. Profound's pricing is not published on the pages reviewed. The Prompting Company publishes self-serve pricing: Basic at $99/mo with 25 tracked prompts, Pro at $299/mo with 100 prompts plus 8 AI-optimized articles, with a free trial before committing.
How quickly will I see results after fixing the friction the benchmark surfaces?
You can retest on a schedule immediately, so a fixed onboarding flow shows up in your next benchmark run. Traffic and mention trends depend on model refresh cycles and indexing behavior, so watch the direction across weeks rather than days.
Do we need to change our product to get started?
No. The platform measures your docs, MCP server, and traffic as they exist today. Improving your score means acting on the friction the benchmarks surface, such as misconfigured setup steps, missing documentation, or unclear error messages, then retesting to confirm the fix landed.
If your MCP server is meant to be wired up by coding agents, every week it goes unmeasured is a week agents route tasks elsewhere. Start a free trial with The Prompting Company and get your first benchmark score this week.