Feedback Week — Wrap-up and Shipped Updates (v4.2.0–v4.3.0)

[saurabhjain1592]

We ran AxonFlow Feedback Week (Feb 6–12, 2026). Wrap-up + shipped updates:

• See the wrap-up comment below
• Full changelog: https://github.com/getaxonflow/axonflow/blob/main/CHANGELOG.md

What to share (anytime)

• What blocked you or slowed you down?
• What was confusing in setup/docs/SDK usage?
• What one improvement would have the biggest impact for your team?
• If nothing blocked you, what feature would make AxonFlow a no-brainer?

Prefer private feedback?

Email hello@getaxonflow.com.

[saurabhjain1592]

Update (Feb 7, 2026): Candidate improvement #1 (source: public community thread)

• Reddit comment: https://www.reddit.com/r/OpenAI/comments/1qqcmfo/comment/o3wrevr/
• Full thread: https://www.reddit.com/r/OpenAI/comments/1qqcmfo/when_openai_calls_cause_side_effects_retries/

What we heard:

• Need live execution status (streaming) while runs are in-flight
• Need a step-level ledger for retries/compensation ("already completed" checks)
• Need clearer guidance on retry/compensation semantics

What we were building / considering next:

• Execution status streaming (SSE or webhooks)
• Step-level ledger with decision + completion + error state
• Retry/compensation guidance for "already completed" behavior

Shipped in v4.2.0–v4.3.0; see the wrap-up comment + CHANGELOG:
https://github.com/getaxonflow/axonflow/blob/main/CHANGELOG.md

[saurabhjain1592]

Closing the loop on AxonFlow Feedback Week (Feb 6–12, 2026). We shipped updates aligned with the themes called out in this thread, especially the “Candidate improvement #1” update (live execution status + step-level ledger for retries/compensation).

What we heard

• Need live execution status (streaming) while runs are in flight.
• Need a step-level ledger for retries/compensation (“already completed” checks).
• Need clearer retry/compensation semantics and guidance.

What we shipped (highlights across v4.2.0 and v4.3.0)

• Live execution status streaming for in-flight runs.
• Step-level execution tracking/ledger improvements to support conditional retries and “already completed” behavior.
• Clearer docs and APIs around retry/compensation semantics.
• Evaluation tier rollout (same features, higher capacity), plus new v4.3.0 limits: cost estimates/day and pending execution approvals.

Full details: https://github.com/getaxonflow/axonflow/blob/main/CHANGELOG.md

If you’re using MAP/WCP or any of the governance features in evaluation or production, reply here with your top blocker (1–2 sentences). Prefer private? Email hello@getaxonflow.com.

[saurabhjain1592]

Closing this thread since Feedback Week has concluded and there hasn’t been follow-up activity. Thanks to everyone who read along. If you hit a blocker or have product feedback, please open a new discussion or issue so we can track it cleanly.