Top Line Goal: Gather performance metrics from AutoRetrieve to provide insight into retrieval deal success

Second Line Goal: Deploy production-grade AutoRetrieve node(s) to make paid Filecoin retrievals accessible via the IPFS Gateways Timeline: 1-2 months

Scope:

• This plan covers only the technological aspects a deployment, not community engagement
• This is a broad overview only. Detailed issue tracking found in retrieval epic of the Project Lightning project board — https://github.com/orgs/filecoin-project/projects/30/views/28

Current state of AutoRetrieve:

• We have an Amazon EC2 instance up running autoretrieve directly on the VM. It is making retrievals and rapidly filling up space
• Application Research is running an instance of Autoretrieve that talks to Estuary’s API

Immediate Decisions:

• Maintain single Github project for Autoretrieve under application research
• Eventually merge the running instances of Autoretrieve so we have a single instance for now

Team (it’s a collaboration of Bedrock and ARG!):

• Hannah - expert for issues relating to the retrieval stack itself, overall tech guidance from Bedrock
• DVD - TPM’ing and day to day management of team
• Kyle - Bedrock eng focusing on ops
• Elijah - ARG eng focusing on autoretrieve software
• Will - overall tech guidance from Bedrock, backup for Hannah
• Why - overall tech guidance from ARG
• Gabe - ARG eng — assisting with autoretrieve

Priorities: