An AI visibility API for your own code
The Proofsource API gives developers and data teams the same numbers as the app as JSON: mention rates, citations, sentiment, query fan-outs, competitors and daily reports from every engine your plan covers. Write keys can also plan and start fixes. The contract is a published OpenAPI spec, so you can generate a client in any language.
curl -s -G https://api.proofsource.co/v2/reports/visibility \
-d brand_id=brd_driffle -d from=2026-10-01 -d to=2026-10-31 \
-H "Authorization: Bearer $PROOFSOURCE_API_KEY"
{
"brand": "Driffle",
"mention_rate": 0.31, "low": 0.27, "high": 0.36, "answers": 412,
"by_engine": [
{"engine": "chatgpt", "mention_rate": 0.38, "answers": 104},
{"engine": "perplexity", "mention_rate": 0.34, "answers": 102},
{"engine": "google_ai_overviews", "mention_rate": 0.29, "answers": 103},
{"engine": "claude", "mention_rate": 0.22, "answers": 103}
]
}AI visibility API, in short
The facts, ready to quote.
- Base URL
- api.proofsource.co, versioned reads under /v2
- Spec
- OpenAPI document at api.proofsource.co/v2/openapi.json
- Reports
- Visibility, citations, sentiment, fan-outs and daily reports
- Screens
- Brands, competitors, citations, questions, fixes and results
- Auth
- Read or write API keys, sent as a bearer token
- Safe retries
- Writes accept an Idempotency-Key, remembered for 24 hours
- Errors
- application/problem+json with a stable type to branch on
- Paging
- List endpoints use cursor paging, so pages never skip or repeat rows
What is an AI visibility API?
An AI visibility API is a programmatic way to get data about how AI answer engines treat a brand: how often it is mentioned, where it ranks in the answer, which competitors appear, which sources are cited and how the tone reads. Teams use it to put AI search next to SEO and revenue data in a warehouse, to build their own dashboards, or to trigger work when something moves.
What makes such an API useful is the same thing that makes the dashboard useful: the data has to be measured on real answers, with enough samples to trust, and it has to be stable enough to build on. A published spec, versioned paths and predictable errors matter as much as the numbers.
The Proofsource API serves the reports and screens of the app over REST under a versioned path, documented by an OpenAPI spec that is checked against the real routes. Read keys pull data; write keys can also plan a fix, see the plan and start exactly that plan, so automation follows the same Yes-first rule as the app.
How it works
A REST API with an OpenAPI spec for AI visibility, citations, sentiment, competitors and fixes, with read and write API keys.
Create a key
In Settings, then API Keys, create a read key to pull data or a write key to also plan and start fixes. It is shown once.
Read the spec
Fetch the OpenAPI document from the API and generate a typed client, or call the endpoints directly.
Pull reports
GET /v2/reports/visibility, citations, sentiment, fanouts or daily for a brand and date range.
Plan and start fixes
POST a plain-words ask to get a plan back with nothing started, then send its plan_id to start exactly that plan.
Retry safely
Send an Idempotency-Key on writes; a repeated request replays the first response instead of acting twice.
See it in the product
Screens from the Proofsource app, drawn with sample data for an example brand.
| Name | Scope | Created by | Last used | |
|---|---|---|---|---|
| Warehouse sync | read | data@driffle.ai | Today, 6:00 AM | Revoke |
| Fix automation | write | dev@driffle.ai | Yesterday | Revoke |
 Looker Studio | read | maya@driffle.ai | 1 hour ago | Revoke |
| Agency rollup | read | ops@driffle.ai | 3 days ago | Revoke |
| Method | Path | Returns |
|---|---|---|
| GET | /v2/reports/visibility | Mention rates by engine or question |
| GET | /v2/reports/citations | Cited sources and where you rank |
| GET | /v2/reports/sentiment | Tone of mentions |
| GET | /v2/reports/fanouts | Searches engines ran behind answers |
| GET | /v2/reports/daily | One row per day and engine |
| GET | /v2/brands, /v2/competitors, /v2/questions | Screen data |
Every capability, live today
9 shipped capabilities behind AI visibility API, each in the product now.
Versioned report API
Read visibility, citations, sentiment, query fan-outs and daily reports under a stable /v2 path.
Screen and fix reads
Read the data behind the brands, competitors, citations and questions screens, and your fixes and their results, as JSON.
Plan and start fixes
Create fixes through the API in two calls: get a plan with nothing started, then start exactly that plan.
Self-service API keys
Create, list and revoke your own keys in the app, each with a read or write scope. A key is shown once and acts with the role of the person who owns it.
OpenAPI contract
A public OpenAPI spec, checked against the real routes so the document and the API do not drift, served from the API itself and ready for code generators.
Idempotent brand creation
A repeated brand create replays the first result instead of making a second brand, so retries are safe.
All client workspaces in one call
For agencies, one call returns every client workspace the agency can access, instead of looping over each.
API cookbook and docs
Public docs cover scans, engines, export, the API, MCP and fix-it, plus an API cookbook of recipes for common developer tasks.
Stable API paging
List endpoints use cursor paging, so pages never skip or repeat rows.
Who it is for
Load AI search into the warehouse
Pull daily reports each morning and join AI mention rates with web analytics and revenue.
Build your own dashboards
Chart visibility and citations next to SEO metrics in the tools your company already uses.
Start fixes from your pipeline
Plan a fix when a page ships, review the plan, then start it and read the proof when it lands.
Report across every client
List all client workspaces in one call and pull each one's reports into your client portal.
Why it matters
A contract you can trust
The OpenAPI spec is checked against the real routes, so generated clients match the API.
Numbers with their samples
Rates come with answer counts and ranges, so downstream charts do not over-read noise.
Yes-first automation
Fixes start only from a plan you have seen, and read keys can never write.
Retries without duplicates
Idempotency keys make timeouts harmless.
Google AI OverviewsGoogle AI Mode
AI visibility API: Proofsource and other tools
Most AI visibility platforms now offer an API. Here is what each one lists on its own pages, and how Proofsource approaches the same job.
| Capability | Proofsource | Profound | Peec AI | Otterly.AI | Scrunch | Rankscale | SE Ranking |
|---|---|---|---|---|---|---|---|
| REST API for AI visibility data | YesVersioned /v2 reports and screen reads. | YesPull Profound data into your own apps, reports and pipelinesSource | YesActions, content, shopping and brand perception endpointsSource | YesBrand reports, prompts, citations and workspace dataSource | YesEnterprise Data API for internal systems, dashboards and AI workflowsSource | YesFull REST APISource | YesAI Results Tracker and SE Visible APIsSource |
| Write actions over the API | YesPlan a fix, then start exactly that plan. | RelatedAgent API runs Profound Agents from external toolsSource | Not listed | YesAdd or delete prompts and tags; manage brand reportsSource | Not listed | Not listed | RelatedSE Visible API covers nearly all UI except billingSource |
| Developer tooling | YesPublished OpenAPI spec checked against live routes. | RelatedSDKs and an API cookbookSource | Not listed | Not listed | Not listed | Not listed | RelatedJSON with a Postman collectionSource |
| Idempotent writes | YesIdempotency-Key on every write, kept 24 hours. | Not listed | Not listed | Not listed | Not listed | Not listed | Not listed |
| API covers what the app shows | YesRead the data behind the brands, competitors, citations and questions screens as JSON | Not listed | Not listed | RelatedBrand reports, prompts, citations, workspace data; monthly request capsSource | Not listed | Not listed | YesProjects, brands, topics, prompts, sources, engines; covers nearly all UI except billingSource |
| API docs and example recipes | YesPublic docs cover scans, engines, export, the API and fix-it, plus an API cookbook | YesReady-made API queries for visibility, citations, metricsSource | Not listed | Not listed | Not listed | Not listed | RelatedUp to 10 requests per second; JSON; Postman collection, no SDKsSource |
Where each of them shines
A mature REST API with SDKs, a cookbook of ready queries and a separate Prompt Volumes API. Compare in full
A wide API that also covers shopping, content and brand perception data. Compare in full
A public API with write access for prompts, tags and brand reports. Compare in full
An enterprise Data API built to feed internal systems and AI workflows. Compare in full
A full REST API with metrics and share links on higher plans.
Several AI search APIs plus a broad SEO API, with clear rate limits and a Postman collection.
Questions about AI visibility API
Does Proofsource have a public API?
Yes. The REST API at api.proofsource.co serves versioned reads under /v2, documented by an OpenAPI spec at api.proofsource.co/v2/openapi.json.
What data can I pull?
Visibility, citations, sentiment, query fan-outs and daily reports, plus the data behind the brands, competitors, citations, questions, fixes and results screens.
Which AI engines does the API cover?
Up to 10 AI engines: ChatGPT, Perplexity, Claude, Google AI Overviews, Google AI Mode, Gemini, Microsoft Copilot, Grok, DeepSeek and Meta AI, depending on your plan. Each report can be broken down by engine.
How do API keys work?
Create them in Settings, then API Keys. A read key can only look; a write key can also plan and start fixes. A key acts with its owner's role, and removing a member revokes every key they own.
Can I start a fix through the API?
Yes, in two calls. The first sends what you want fixed and returns a plan with nothing started. The second sends that plan_id and starts exactly that plan.
Is it safe to retry a request that timed out?
Yes. Send the same Idempotency-Key and you get the first response back instead of a second action. Keys are remembered for 24 hours.
Can an agency read every client at once?
Yes. One call lists every client workspace the agency can access, so you do not loop over each client yourself.
How is this different from the Profound or SE Ranking APIs?
Both offer strong APIs, Profound with SDKs and a cookbook, SE Ranking alongside a large SEO API. The Proofsource API is focused on measured answers for your own buyer questions with sample sizes, and on a plan-then-start fix flow you can drive from code.
Keep exploring
Try it on your own brand.
Proofsource is live: scan your brand and see AI visibility API at work with your own answers, competitors and sources.
Google AI OverviewsGoogle AI Mode