1. Use a tool: List the latest runs
  2. Ask the model: Give the results to the model
  3. Use a tool: Go through each run
  4. Use a tool: Read its turns
  5. Ask the model: Give the results to the model
  6. Ask the model: Summarize them
  7. If / otherwise: Anything sensitive?
  8. Ask a person: Ask the owner
  9. Use a tool: Post the digest
  10. Ask the model: Give the results to the model

Recent runs digest

Summarizes the latest 20 runs of whoever starts it and asks before including anything sensitive.

OperationsSchedule · Slack11 stepsFrom Momor · read-only

Use this workflow →
Read-only · sample data

Press “Try a sample run” to watch it work with sample data. Nothing real is touched.

How it works

  1. 01Every eveningSchedule
  2. 02List the latest runsThe caller’s own, at most 20
  3. 03Give the results to the modelThe list of runs
  4. 04Go through each runOne at a time
  5. 05Read its turnsTheir state, request and result
  6. 06Give the results to the modelThe turns
  7. 07Summarize themWhat changed, what is waiting
  8. 08Anything sensitive?Yes asks first
  9. 09Ask the ownerInclude this?
  10. 10Post the digestSlack, without anything sensitive if the owner declines
  11. 11Give the results to the modelIt writes the run’s result

What it connects to over MCP

  • SlackPosts the digest
  • The digest ownerDecides on anything sensitive

Gateway picks a model for each AI step. You can change it in the Designer.

After Momor drafts it

  1. Connect the tools it uses.
  2. Choose who gets asked.
  3. Test run it with your own data. A test run saves the workflow as a version, then runs it for real.

Make it yours.

Open this template as a new draft in the Designer and change it to fit your process, or bring us the process and we build it with you.