Published on March 4, 2026
Sellm Opens Its Platform: Full API Access for AI Search Tracking
Now available: REST API access for AI visibility tracking across ChatGPT, Claude, Perplexity, Gemini, and Grok.
Starting today, everything you can do in the Sellm dashboard is available through a public REST API. That means teams can pull AI search visibility data directly into their own dashboards, reporting pipelines, alerting systems, and internal tooling - without relying on manual exports or screenshot workflows.
Why We Built This
AI search is becoming a meaningful acquisition channel. When someone asks ChatGPT for "the best CRM for startups" or Perplexity for "top project management tools," the brands that get mentioned win real traffic and trust. But tracking that visibility has been difficult to integrate into existing workflows.
Teams already have BI tools, data warehouses, reporting cadences, and alerting infrastructure. What they needed was a reliable, structured data source for AI visibility - not another dashboard to check. That's what the API delivers.
What the API Covers
The API exposes the same data and capabilities as the Sellm dashboard, organized around a few core areas:
Analysis and Results
The heart of the platform. You can list analysis runs, pull run-level KPI summaries (share of voice, coverage, average position, sentiment), and drill into per-prompt and per-provider breakdowns. If your brand ranks #2 in ChatGPT for a specific query but #5 in Claude, you'll see that distinction clearly in the API response.
- Run KPI summaries: share of voice, coverage, average position, sentiment
- Per-prompt breakdowns with individual scores, competitors, and opportunities
- Per-provider breakdowns across ChatGPT, Claude, Perplexity, Gemini, and Grok
- Historical trend data across your most recent completed runs
Prompt Management
Prompts are the queries that get sent to AI providers during analysis. The API lets you create, update, delete, and batch-manage prompts without opening the dashboard. You can assign specific providers to each prompt - run some queries only against ChatGPT and Claude, others across all five platforms.
Project and Settings
Retrieve project details, configure competitors, manage prompt tags for filtering, and control scheduled runs. You can enable or disable weekly automated runs, and trigger manual analysis runs when you need fresh data outside the regular schedule.
Supported Providers
The API returns data from every AI search platform that Sellm tracks. As of today, that includes:
- ChatGPT (OpenAI) - The most widely used conversational AI, handling hundreds of millions of queries daily. ChatGPT tracking tells you how OpenAI's models position your brand in buying and research contexts.
- Claude (Anthropic) - Popular among enterprise teams and analytical professionals. Claude tracking captures how your brand appears in longer, more considered AI responses.
- Perplexity - An AI-powered search engine that cites sources directly. Perplexity tracking reveals whether your content gets cited and where you rank in source-backed answers.
- Gemini (Google) - Google's AI model powering search and assistant features. Gemini tracking shows how Google's AI layer treats your brand, which matters for users embedded in Google's ecosystem.
- Grok (xAI) - Built by xAI and integrated into X (formerly Twitter). Grok tracking captures your brand's presence in this growing platform with its distinct user base.
Every endpoint accepts a provider filter, so you can isolate results for any single platform or compare them side by side.
How Teams Are Using It
During the early access period, we've seen a few common integration patterns:
- Weekly reporting automation. Agencies pulling AI visibility data per client into Looker, Tableau, or Google Sheets alongside SEO and paid metrics.
- Data warehouse pipelines. Growth teams streaming analysis results into Snowflake or BigQuery, joining AI visibility data with conversion funnels and CRM data.
- Slack and email alerts. Engineering teams building monitors that fire when share of voice drops below a threshold or a competitor overtakes their position in ChatGPT or Claude.
- Internal dashboards. Product marketing teams building branded dashboards that combine AI visibility data with web analytics for executive reporting.
Authentication and Getting Started
The API uses project-scoped API keys with bearer token authentication. You can create and rotate keys in the dashboard without downtime. A typical first integration looks like:
- Create a project and configure your prompts (in the dashboard or via the API)
- Generate an API key in the project settings
- Make your first authenticated request to
GET /v1/project - Wait for a scheduled or manual analysis run to complete
- Pull results from
GET /v1/analysis/runs/{'{'}runId{'}'}/summary
The full API reference, quickstart guide, and authentication docs are available at sellm.io/docs/api.
What's Next
This release covers the core platform functionality. We're working on webhook support for run-completion notifications, dedicated endpoints for competitive intelligence, and expanded provider coverage as new AI search platforms emerge.
If you're already a Sellm user, you can generate an API key today in your project settings. If you're new, create a free account to get started - the free tier includes API access with the same capabilities as the dashboard.
Frequently Asked Questions
Is the API available on the free plan?
Yes. The API is available on all plans, including the free tier. Free accounts have the same prompt and run limits that apply in the dashboard.
Can I track ChatGPT and Claude separately?
Yes. Every results endpoint accepts a provider filter. You can pull data for ChatGPT alone, Claude alone, or any combination of providers.
What are the rate limits?
Read endpoints have generous limits for dashboard-style polling. Manual analysis runs are limited to 1 per 7-day window per project to prevent abuse. Rate limit headers are included in every response.
Do I need to use the dashboard at all?
You need the dashboard to create your first project and generate an API key. After that, everything - prompt management, run triggering, data retrieval - can be done entirely through the API.