Conversation
A reasoning model's thinking (OpenAI reasoning deltas, reasoning_content from OpenAI-compatible providers such as DeepSeek) already reached the agent as llm.ReasoningTextDelta and stopped there. A written reply now reports it as an agent.ReasoningDelta, and a persistent conversation carries it to watchers as a top-level "reasoning" string on the reply's ephemeral Chat updates, keeping only its last 4000 bytes and riding the existing live-update tick. It never reaches the outbox, the local ledger or the settled message, and it is reset when the next command begins. Thinking can name tools and internal details, so it is shown only for an agent config with the new show_reasoning turned on. It is off by default, including for every existing config.
Nash0x7E2
force-pushed
the
nash/conversation-artifacts
branch
from
October 1, 2026 17:54
2f6806f to
bfd4bce
Compare
Nash0x7E2
force-pushed
the
nash/conversation-live-reasoning
branch
from
October 1, 2026 17:55
70a73b6 to
dbe68ff
Compare
Member
Author
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Why
Athena web shows a "Thinking…" disclosure on a reply while it is being written. It reads a top-level
reasoningstring off the Chat message, which only arrives on ephemeralmessage.updatedupdates; it keeps what it saw in memory and forgets it on reload. The stale Athena runtime branch (2d40a966) wrote that field. #699 ported that commit's tool steps but not the reasoning. On accelerate the LLM router already streams reasoning asllm.ReasoningTextDelta(OpenAI reasoning deltas,reasoning_contentfrom OpenAI-compatible providers such as DeepSeek), but the agent dropped it.Reasoning can name tools and internal details, so it is opt-in: an agent config has a new
show_reasoningboolean, off by default, and every existing config starts off.Changes
agent.ReasoningDelta. A voice never gets it.show_reasoningon keeps the last 4000 bytes of the current reply's thinking and adds them asreasoningto the ephemeral update it already sends on its live tick. Reasoning never goes into the outbox, the local ledger or the settled message, and it resets when the next command starts.show_reasoningis added toAgentConfig,AgentConfigRequestandAgentConfigPatch, with migration20261001130000, a regenerated spec, the Go client and JS types, and a note in the sdk skill for the other SDKs.🤖 Generated with Claude Code