Zurück

How to optimize your token use?

Understanding how to get the most out of Astell without burning through your tokens

Why optimize token usage?

Tokens are only consumed by Processed Data ingestion, not by chat. Optimizing token usage isn't about limiting conversations—it's about being intentional with what content you choose to process so tokens are spent on information that actually creates value.

Roughly how far do tokens go on each plan?

Although Astell's proprietary ingestion dramatically lowers ingestion costs, each plan is designed to support different volumes of data ingestion. For details, see Real-world usage examples

How can I reduce token usage for document processing?

Biggest savings come from processing less junk and focusing on what matters:

• Use manual processing (turn off auto-sync; upload key documents only when needed)
• Limit the date range (sync last 6–12 months; process older files only when needed)
• Review integrations quarterly (remove inactive sources; tighten selective sync over time)
  We are also rolling out our selective folder sync feature to enable users to sync active work folders and exclude archives/downloads/personal folders.

How do I make sure all my data can be ingested within the token limit?

If you're onboarding a large workspace or importing historical files, the first month often requires a lot of tokens because Processed Data ingestion (documents, images, audio, video, meetings, and web pages) is what consumes tokens.
To avoid ingestion slowing down or pausing mid-onboarding, plan up front for how you want overage handled. If you want the initial ingestion to complete without stopping, set a higher token limit in your first month so ingestion can continue beyond your included monthly pool via pay-as-you-go. Any pay-as-you-go tokens you use are billed on your next billing date at your plan's per‑1,000 token rate.
After the initial ingestion finishes, token usage usually drops significantly because:

• Native Data continues syncing for free
• Only new or changed Processed Data consumes tokens
• You can tighten selective sync rules once the initial backlog is done
  For pay-as-you-go rates and how token limits work, see “Pay-as-you-go, token limits, and upgrades”.

How often should I review token usage and plan fit?

Review every 3 months: check average consumption, identify optimization opportunities, confirm your plan still fits, and update team policies based on real usage.

How do I get help with tokens if I'm stuck?

If you need help, email founders@labtwofour.com with subject “Token Usage Assistance” and include your usage patterns plus the questions you're trying to solve. Enterprise customers can also use dedicated support (quarterly usage reviews and custom optimization consulting).

How can I reduce token usage with selective syncing?

Selective syncing keeps the important data flowing while avoiding unnecessary processing. You can currently set date ranges (like the last 12 months). In the future, you will be able to limit channels/folders/repos and exclude certain attachment types. See "How to monitor your token usage"