Skip to content
Opusgen
Get free report
The free Agent Report for developer tools See a sample

See what AI agents do with your product.

Opusgen tracks the coding agents that read your docs, sign up and call your API. See where they fail and what it costs you, then prove each fix in a sandbox before you ship.

Get your free Agent ReportExplore the demo

Free for developer tools. The report needs no integration.

PostboxOverview

Illustrative data
Agent sessions
48,210+31%
Agent share of signups
58%+9 pts
First working call
43%−6 pts
Agent-sourced MRR
$38.4k+$6.1k

Requests per day

  • Agents
  • Humans
3k
6k
9k
Agents pass humans
Sep 10Sep 20Sep 30Oct 9

Top failure

403 on first send

38% of failed agent tasks

POST /v1/send

Failed tasks by cause

  • 403 on first send38%
  • Deprecated SDK call24%
  • MCP schema mismatch14%

Fix ready · #214

Live

Times in UTC
  1. claude-code POST /signup201
  2. gemini-cli tools/call send_emailok
  3. cursor POST /v1/send403
  4. cursor POST /v1/send403
  5. codex POST /v1/api-keys201
  6. claude-code GET /docs/quickstart.md200

Identifies the agents your users already run

Every request is matched to the agent that sent it, by cryptographic signature, MCP client name, user agent or SDK tag. Each label carries a confidence score.

identify.log

Illustrative data
  • 14:02:11agent=claude-codeconf=0.95
  • 14:02:12agent=codexconf=0.96
  • 14:02:14agent=cursorconf=0.97
  • 14:02:15agent=gemini-cliconf=0.91
  • 14:02:17agent=chatgpt-agentconf=0.99
  • 14:02:18agent=mcp-clientconf=0.84

Your dashboard logged a signup. The agent installed a competitor.

Agents don't click, scroll or file tickets. They read your docs as text, call your API, hit an error and switch vendors in a few minutes. Click-based analytics records one healthy signup and nothing else.

Your product analytics

New user · 14:02

1 new signup

Page views
0
Clicks
0
Session length
0:00
Support tickets
0

Events since signup

No events recorded

Account statusHealthy

What the agent actually did

cursor · 4 min 12 s

  1. 0:00Reads four docs pages as MarkdownGET /docs/quickstart.md
  2. 0:41Creates an account through your APIPOST /v1/accounts · 201
  3. 0:58First call returns 403 ForbiddenPOST /v1/send · 403
  4. 1:06Retries four times, same errorPOST /v1/send · 403 ×4
  5. 2:30Installs a competitor insteadnpm install relaymail
  6. 4:12Tells its user: "Done."Session ends

1 customer lost

Your dashboard still says: healthy.

Illustrative session. Relaymail is a fictional vendor.Replay this session in the demo

Agents became the majority user of developer software in 2026.

It took six months and reached every industry in Mintlify's data. Most of these visitors never render a page, so a pageview-based dashboard counts almost none of them.

  • 66%

    of traffic on Mintlify-hosted docs came from agents in July 2026, up from 15% in January1

  • 213M

    agent requests against 105M human page loads on Mintlify-hosted docs in July 20261

  • 70%

    of new Supabase databases are created by agents or AI-driven tools2

  • 80%+

    of databases provisioned on Neon were created by AI agents3

Fig. 1Agent share of docs traffic, by industry

  • Feb 2026
  • Aug 2026
Agent share of docs traffic by industry, February and August 2026, in percent
IndustryFebruary 2026August 2026
Fintech & payments20.5%70%
Crypto & web324.6%62.5%
Agent platforms & AI tooling13%47.3%
AI coding tools12.3%46.8%
AI models & inference9.1%41.4%
Dev infrastructure & data15.5%37.9%
SaaS & business apps14.4%34.4%

Agent share of docs traffic on Mintlify-hosted docs, by industry, February to August 2026.4

The work has a name now. Microsoft defined agent experience (AX) by two measures: whether an agent chooses you, and whether it uses you correctly. Netlify open-sourced AXIS to score it.56

Agents are invisible in production, but perfectly testable in a sandbox.

Opusgen runs two engines. A production pipeline records what agents did on your docs, API, MCP server and signup. An agent lab runs real coding agents against you every night to show why. Both feed one diagnosis.

Production pipeline

What agents did

  • Docs
  • API
  • MCP
  • Signup
  • Billing

Agent lab

Why they did it

  • 5 agents
  • 50 tasks
  • Nightly sandboxes

Diagnosis

  1. Failure clusters
  2. Fix pull requests
  3. Verified lift

Production pipeline

Know which agent sent every request.

Add a few lines at your docs edge, API gateway, MCP server and signup flow. Signed agents, MCP client names, declared user agents and SDK tags tell us which agent sent each request, and how sure we are.

  • Cryptographic agent signatures (Web Bot Auth) are verified7
  • MCP client names and SDK tags are read on every call
  • Every label carries a confidence score
14:02:14POST/v1/api-keys201

Signals

  • Agent signatureverified
  • Declared user agentclaude-code
  • SDK tagpostbox-node 4.2
  • Docs read as Markdown3 s before call

Identified as

claude-code

Harness
terminal
Surface
api
Account
acct_8f2c
Confidence0.94
Illustrative dataOpen in the demo

Production pipeline

Follow every agent-started account to revenue.

We stitch docs reads, signups, keys and calls into one journey per account, and keep the agent attribution all the way through billing. Split the funnel by agent and see where each one drops.

Leadership gets one number for the board: agent-sourced revenue.

Agent funnel

Agent-started accounts · last 30 days

  1. Docs read4,210
  2. Signup1,860
  3. API key1,420
  4. First working call61043% of keys reach a working call
  5. Paid92
Agent-sourced MRR$38.4k
Illustrative dataOpen in the demo

Agent lab

Find out why they fail before your customers do.

Every night, real coding agents run realistic tasks in sandboxes against you and your category. We grade whether the agent picked you, whether the code worked, and how many turns, tokens and minutes it took. Failures cluster by cause and point at the page, error or tool behind them.

Nightly lab run

Run 412 · 40 tasks · 5 agents

12345678
claude-code
codex
cursor
gemini-cli
chatgpt-agent
PassFail1 install · 2 auth · 3 send · 4 domain · 5 webhook · 6 template · 7 retry · 8 billing
Picked Postbox
18%
Code worked
72%
Median turns
14
Median time
4.2 min

Failure clusters

  1. 403 on first send38%
    → src/errors.ts
  2. Deprecated SDK call24%
    → /docs/quickstart
  3. MCP schema mismatch14%
    → tools/send_email
Illustrative dataOpen in the demo

postbox/api #214

Name the missing scope in 403 errors

opusgen[bot]

src/errors.ts

48throw new ApiError(403, "Forbidden");
48throw new ApiError(403, "This key can't send email. Add the send:email scope at /settings/keys.");

Opusgen lab· 40 runs on previewPassed

Lab success rate+18 pts

main72%
preview90%
Illustrative dataOpen in the demo

Diagnosis

Ship the fix with the proof attached.

Opusgen opens a pull request against your docs, error messages or MCP tool descriptions. It reruns the lab on your preview deploy and posts the lift before you merge.

  • Model-release alerts

    A new model ships overnight. By morning you know if agents still succeed with you.

    Illustrative data

    Model release06:12

    codex vNext: efficacy −12 pts on /v1/send

  • Warehouse export

    Agent events and revenue land in Snowflake or BigQuery next to your own data.

    Illustrative data

    agent_events

    ts
    timestamp
    agent
    string
    surface
    string
    account_id
    string
    revenue_usd
    numeric

Find out if agents pick you.

The free Agent Report runs real coding agents on 50 tasks in your category, across five popular stacks. You see who they choose, how often they succeed with you, why they skipped you, and the one fix that would move you most.

Agent Report · Email APIs

50 tasks · 5 stacks · 5 agents

Share of agent picks

Tasks that didn't name a vendor

  1. Relaymail41%
  2. Sendwell27%
  3. Postbox (you)18%
  4. Letterkit14%

Success when named

Tasks that asked for you by name

64%

Why agents skipped you

  1. Quickstart calls a deprecated send() signature34% of runs
  2. 403 on first send with the default key scope29% of runs
  3. Docs served as HTML only, so agents ran out of context18% of runs

The fix that would move you most

Return a 403 that names the missing scope and links to the fix. Lab rerun: +9 pts pick share.

Sample report · fictional vendors and numbers

Live in an afternoon.

Three small changes on your side. No client-side script, and nothing your users ever see.

  1. 1Add the edge snippetOne middleware in front of your docs and API.
  2. 2Wrap your MCP serverTool calls arrive with the client's name attached.
  3. 3Connect billingAgent-started accounts carry through to revenue.
proxy.ts
// proxy.ts
import { opusgen } from "@opusgen/next";

export default opusgen({ key: process.env.OPUSGEN_KEY });

What we see, and what we never see

We see

  • Requests to your docs, API, MCP server and signup
  • Which agent sent each one, and how sure we are
  • The account and billing events you connect

We never see

  • Prompts, or the conversation between a person and their agent
  • PII. It is redacted at ingest, and you control retention

Not owned by any model provider. Every agent is measured the same way. [SECURITY_STATUS]

Start with a free report.

Pricing is in beta.

  • Free

    $0

    For: Seeing where you stand

    Includes

    • Monthly Agent Report for one category
    Get free report

  • Growth

    Most teams start here

    from [GROWTH_PRICE]/mo

    For: Teams with live agent traffic

    Includes

    • Production telemetry
    • Weekly lab runs
    • Alerts
    Talk to us

  • Scale

    Custom

    For: Companies where agents drive revenue

    Includes

    • Nightly lab
    • Benchmarks
    • Fix pull requests
    • Warehouse export
    Talk to us

  • Enterprise

    Custom

    For: Multiple products

    Includes

    • Custom task suites
    • SSO
    • Data residency
    Talk to us

Common questions

Not covered here? Email us [CONTACT_EMAIL]

See your product the way agents do.

Tell us where to look. We'll run the agents and send your report.

Your report shows

  • Your pick rate against your category
  • Your success rate when agents ask for you by name
  • Your top three failure causes
  • One concrete fix, tested in the lab

By requesting a report you agree to our privacy policy [PRIVACY_URL].