Automatically extract document data into Airtable entries with AI-powered automation. Save time and eliminate manual data entry.
npm install && npm run dev
- Next.js
- TailwindCSS
- Shadcn
- Airtable
- OpenAI
- Drizzle
- Resend
Current architecture assumes you have the following accounts and related keys:
- Airtable as the final destination for the extracted data
- OpenAI to extract data from documents
- Resend to send emails
- Cloudmailin to receive emails
- AWS for hosting files
- Vercel host hosting the frontend
- Neon Database for postgres
My goal is to make this as easy as possible to self-host with a minimum number of external SaaS dependencies to make sure you data doesn't have to leave your own infrastructure.
Plan is to use sst.dev and keep everything in the same AWS account. Neon => Aurora DSQL Resend => SES Cloudmailin => SES Vercel => OpenNext OpenAi => There is no easy way to replace OpenAi but we can allow changing the provider to use Claude through Bedrock or other LLM providers. Airtable => similarly we can export data to other services like Notion, Google Sheets, etc.
Inspired by Plausible, AskAir is open-source under the GNU Affero General Public License Version 3 (AGPLv3) or any later version. You can find it here.