Profound Costs $399/mo for 3 Engines — I Built the Same AI Visibility Tracker on Serpent API (With Code)

Bottom line up front: Profound is the most complete AI visibility platform on the market — a genuine billion-dollar company with hundreds of enterprise customers — and its first multi-engine plan is $399/mo ($4,788/yr), with everything serious priced as custom Enterprise. For a lot of teams that is the right buy. But the core job — “when people ask AI about my category, am I cited?” — is something you can build yourself. I did, on Serpent API’s AI Ranking API, and it tracks brand citations across Claude, ChatGPT, Gemini and Perplexity for cents per check. This post is the honest teardown: what Profound does, what the DIY version does, the working Python code, and exactly where each one wins.

I want to be fair to Profound up front, because this is not a hit piece. It earned its valuation by being genuinely good. The question this post answers is narrower and more useful: if all you need is the core citation tracking, do you need to pay enterprise SaaS prices for it? The answer is no — and the gap is large enough to matter.

What Profound costs, accurately

Let’s get the numbers right, because they changed recently. Profound retired its old $499/mo Lite plan in late 2025. The current published pricing is:

So the first plan that tracks more than ChatGPT is $399/mo, and to cover the full set of assistants (Gemini, Copilot, Claude, Grok and the rest) plus API access and SSO you are into custom Enterprise pricing. That is entirely reasonable for the depth Profound offers — it is just a meaningful number if your actual need is “tell me whether I’m cited across the big four assistants.”

What Profound actually does

To compare fairly you have to respect what Profound is. Its headline capabilities:

• Answer Engine Insights — brand visibility, citations and share of voice across ChatGPT, Perplexity, Google AI Overviews, Gemini, Copilot, Claude, Grok and more.
• Conversation Explorer — real-time data on what people actually ask AI in your category (a kind of AI “search volume”).
• Agent Analytics — bot-traffic monitoring (GPTBot, ClaudeBot, PerplexityBot and friends) from your server and CDN logs.
• Sentiment and competitor benchmarking — how you are described, and how you stack up.
• A polished, multi-seat dashboard with exports, and API access on Enterprise.

That is a lot, and most of it is genuinely useful. A DIY build does not replicate all of it — and I am not going to pretend it does. What a DIY build replicates is the part most people are actually paying for: the citation tracking and share of voice.

Feature parity: Profound vs DIY on an API

Read honestly, the table says: Profound wins on breadth, polish and the two features a DIY build genuinely can’t cheaply replicate (Conversation Explorer and bot-traffic analytics). DIY wins decisively on cost, data ownership and native API access for the core citation job. If the four rows that matter to you are the first four, you are paying $399/mo for something you can build in an afternoon.

The build, step by step

Here is the whole thing. It uses Serpent API’s AI Ranking API, which takes a prompt and a brand and returns, for each assistant, whether the brand was cited, at what position, and the sources that assistant used.

Step 1 — define what to track

PROMPTS = [
    "best project management software for agencies",
    "top alternatives to asana",
    "which pm tool is best for remote teams",
]
BRAND = "YourBrand"
COMPETITORS = ["Asana", "Monday", "ClickUp"]
API_KEY = "sk_live_..."   # from your Serpent API dashboard

Step 2 — query all four assistants for one prompt

The combined endpoint checks every assistant in a single call and returns a clean, structured result — including the public model name for each engine.

import requests

def check(prompt):
    r = requests.get("https://apiserpent.com/api/ai/rank/all", params={
        "prompt": prompt,
        "brand": BRAND,
        "api_key": API_KEY,
    }, timeout=120)
    r.raise_for_status()
    return r.json()["results"]   # one entry per assistant

# Example shape of one entry:
# {
#   "model": "claude-sonnet-4-6",   # clean public model name
#   "cited": true,
#   "position": 2,
#   "sources": ["https://...", "https://..."]
# }

Step 3 — store the results

import sqlite3, datetime

db = sqlite3.connect("ai_visibility.db")
db.execute("""CREATE TABLE IF NOT EXISTS checks(
    day TEXT, prompt TEXT, model TEXT, cited INT, position INT)""")

def store(prompt, results):
    day = datetime.date.today().isoformat()
    for e in results:
        db.execute("INSERT INTO checks VALUES (?,?,?,?,?)",
            (day, prompt, e["model"], int(e["cited"]), e.get("position")))
    db.commit()

for p in PROMPTS:
    store(p, check(p))

Step 4 — compute share of voice

# What share of (prompt x assistant) checks cited you today?
rows = db.execute("""SELECT model,
    SUM(cited)*1.0/COUNT(*) AS visibility
    FROM checks WHERE day = ? GROUP BY model""",
    (datetime.date.today().isoformat(),)).fetchall()

for model, visibility in rows:
    print(f"{model:20s} {visibility:.0%} of prompts cited you")
# claude-sonnet-4-6    67% of prompts cited you
# gpt-5.5              33% of prompts cited you
# gemini-3.1-pro       33% of prompts cited you
# sonar-pro            67% of prompts cited you

Step 5 — schedule it

Drop the script on a daily cron (or an n8n / GitHub Actions schedule) and you have a running AI visibility tracker. Add the competitor names to the same prompts to chart your share of voice against them, and add a chart with any plotting library. The full version — competitor share of voice, charts and alerts — is in our cross-engine AI share of voice tracker, and the same pattern applied to citations specifically is in our AI citation tracker in Python. For a full front-end, see the brand visibility dashboard tutorial.

The cost gap, worked out

Serpent API’s AI Ranking is pay-as-you-go. A combined check across all four assistants is billed at the 4-engine rate — about 4 cents at the Default rate, dropping toward a fifth of a cent at higher volume. Match Profound Growth’s 100-prompt allowance and run it daily across four assistants:

Even at a daily cadence on the entry Default rate, the DIY build ($120/mo) covers four assistants for less than a third of Profound’s $399/mo for three — and once you have deposited enough to reach the Growth rate, the same daily tracking is about $12/mo. Weekly tracking, which is plenty for most brands, is ~$17/mo at the Default rate. That is a 20×–60× difference for the core citation job. Credits never expire and the minimum deposit is $10, so there is no subscription to waste on a quiet month. (Our pricing page has the live AI Ranking rates.)

When you should just pay for Profound

Buy Profound — or another full platform — if any of these is true:

• You need the Conversation Explorer’s AI “search volume” data, which a DIY build cannot cheaply reproduce.
• You need agent/bot-traffic analytics correlated with visibility.
• You need 10+ engines, sentiment, and a multi-seat dashboard with exports for stakeholders who will never touch a database.
• Your team has no engineering time, and a finished product tomorrow is worth more than the price difference.

If none of those is true — if you mainly need to know, reliably and on a schedule, whether you are cited across the major assistants and the Google AI Overview — the DIY build does that job for a tiny fraction of the cost. For the full landscape of paid tools (Otterly, Peec, Rankscale, Semrush and Profound), see our comparison of 6 AI visibility tools; for the metric underneath all of them, AI visibility scoring explained; and for the AI Overview block specifically, which SERP APIs return Google AI Overviews.

FAQ

How much does Profound cost in 2026?

$99/mo Starter (ChatGPT only), $399/mo Growth (3 engines, 100 prompts), and custom Enterprise for 10+ engines, API and SSO. The old $499/mo Lite plan was retired in late 2025; billing is annual.

Can I build my own alternative to Profound?

Yes. It is scheduled prompts, a query against each assistant for your brand, storage and a share-of-voice calculation. Serpent API’s AI Ranking API returns citation data across Claude, ChatGPT, Gemini and Perplexity per prompt — the working code is above.

What does a DIY AI visibility tracker cost to run?

A combined check across four assistants is about 4 cents at the Default rate, less at scale. 100 prompts daily across four assistants is roughly $120/mo at Default, ~$12/mo at Growth, versus Profound’s flat $399/mo for three engines. Weekly tracking is ~$17/mo. Pay-as-you-go, $10 minimum, credits never expire.

What does Profound do that a DIY tracker doesn’t?

A polished dashboard, Conversation Explorer (AI search volume), agent/bot-traffic analytics, sentiment, out-of-the-box competitor benchmarking and 10+ engines on Enterprise. A DIY build gives you the core citation data across four assistants and the AI Overview cheaply; you build the dashboard and extras.

Which AI engines can the DIY tracker cover?

Four assistants — Claude, ChatGPT, Gemini and Perplexity (clean model names: claude-sonnet-4-6, gpt-5.5, gemini-3.1-pro, sonar-pro) — plus the Google AI Overview block via the SERP API.

Is Profound worth it?

For a large brand needing depth, breadth and a finished dashboard, yes — that is why it is a billion-dollar company. If you mainly need core citation tracking and want to control cost and data, a DIY build does the job for a fraction of the price.