What is Account Connect?
Account Connect brings provider billing and usage data into SpendLens AI. It shows where your OpenAI and Anthropic spend goes and highlights areas worth reviewing.
FAQ
What teams ask before connecting.
Quick answers about Account Connect, setup, privacy, model recommendations, and savings estimates.
Do I need to install an SDK to use Account Connect?
No. Connect OpenAI or upload Anthropic reports to start. Add SDK telemetry later only when you need deeper application-level evidence.
How does the OpenAI connection work?
Add a dedicated OpenAI Admin key for SpendLens AI reporting. SpendLens AI imports organisation costs, projects, API keys, models and token usage automatically.
Will SpendLens AI use my OpenAI key to call a model?
No. The connection makes reporting requests only. It does not send prompts, generate responses or change your OpenAI account.
How is my OpenAI credential protected?
The credential is encrypted before storage and is never shown again. Use a dedicated Admin key so you can revoke it from OpenAI at any time.
Do I need an Anthropic API key?
No. Export the official Cost and Usage CSV reports from Anthropic Console and upload them one at a time. SpendLens AI detects which report you selected.
Which Anthropic reports should I upload?
Upload both Cost and Usage reports for the same date range. Group each report by Model, keep the other filters set to All, and use Service Account for Cost if your organisation needs that spend included.
Does SpendLens AI keep previous Anthropic uploads?
Yes. Report values are stored by date, so uploading August does not remove July. Uploading the same dates again updates matching records instead of intentionally adding duplicates.
Can I connect more than one AI account?
Yes, subject to your plan limit. Free supports 1 connection, Starter 2, Pro 5 and Growth 15. Use the AI account selector to view each account separately or combine them.
What can I see in the Account Connect dashboard?
You can review provider-billed spend, projects, models, token usage, API keys, ownership gaps, caching signals, batch usage and model savings candidates when the provider data supports them.
Why are some Account Connect sections empty?
Providers do not expose every metric for every account or report. Upload both Anthropic reports or refresh OpenAI first. SpendLens AI shows only evidence available for the selected account and date range.
What happens when I delete a connected account?
SpendLens AI permanently removes its stored credential, imported costs, usage, resources, ownership mappings and sync history.
What is the difference between Account Connect and SDK telemetry?
Account Connect shows where the provider bill goes. SDK telemetry connects cost to specific parts of your application and helps validate whether a model change can preserve quality.
Do you proxy my LLM traffic?
No. Your OpenAI and Anthropic calls continue going directly to the provider. The SpendLens SDK sends usage metadata after calls complete.
What is Compare Mode?
The SpendLens SDK runs up to three candidate calls inside your staging application using your provider keys and reports aggregate cost, tokens, speed, reliability and errors. Optional quality evaluation runs separately on SpendLens infrastructure.
Where do Compare Mode tests run?
Candidate model calls run in your staging application with provider keys you control. Optional quality evaluation runs on SpendLens infrastructure with SpendLens provider credentials, a redacted template and safe generated data.
Does Compare Mode read customer prompts?
No dynamic customer values or production responses are used for hosted quality evaluation. SpendLens uses only the reusable redacted template and safe generated examples.
Which models support hosted quality testing?
The current version supports OpenAI, Anthropic, Gemini, DeepSeek, Mistral and Qwen. You can enable or disable providers from the Compare Mode page. More providers are coming soon.
What are Compare credits?
Compare credits cover model calls and judging performed by SpendLens during optional hosted quality evaluation. They do not pay for candidate calls made in your staging application with your provider keys. Pro includes $40 each month and Growth includes $100.
What do I get after installing?
A dashboard showing which projects, workloads, models, and providers are driving AI spend, plus cost-efficient model recommendations to test.
Do you store prompts?
Full user prompts and model responses are not stored by default. Template-only mode may use system prompt or template text for workload classification.
Which providers are supported?
SpendLens currently tracks OpenAI and Anthropic usage. Recommendations can compare suitable models from leading providers.
Are savings guaranteed?
No. Savings are estimates. SpendLens helps you identify opportunities worth validating before changing production workloads.
Do I need to tag every task?
No. Mark the business workflows that matter with the Python decorator or Node.js observe helper. Use client tags when you need more precise per-call attribution.