Repository navigation
docs(blog): polish bidi GA announcement [replica of harness-sdk#4904] - #22
poshinchen wants to merge 2 commits into
Conversation
poshinchen
left a comment
poshinchen
left a comment
There was a problem hiding this comment.
Strandly v3-reason (optimized) review
Assessment: approve
Assessment: Approve
This is a focused, documentation-only PR that improves the clarity and grammar of the Bidi GA announcement. The wording polish, hyphenation, subject-verb agreement fixes, and backtick formatting of BidiAgent are all solid, low-risk improvements, and the new setup-docs pointer meaningfully improves the FAQ.
Review Categories
- Accuracy: A few customer-facing claims rely on external facts not visible in the diff — the new
/docs/user-guide/sdk/bidi/link, vendor model names (GPT-Realtime-2.1, Gemini 3.8 Live, Nova 2.5 Sonic), and the reworded telemetry description. Flagged for verification rather than as blockers. - Quality: Prose edits are clear improvements; no structural or maintainability concerns.
Nice cleanup that makes the announcement read more smoothly and gives readers a concrete next step.
| ### What does general availability mean for the API? | ||
|
|
||
| The supported Bidi API follows the SDK's versioning and deprecation policy. Our latest docs will show how to be set up with the latest. | ||
| The supported Bidi API follows the SDK's versioning and deprecation policy. See the [bidirectional streaming docs](/docs/user-guide/sdk/bidi/) for setup instructions and supported model providers. |
There was a problem hiding this comment.
Important
Issue: The new link /docs/user-guide/sdk/bidi/ is introduced here; a broken docs link would undermine the "useful pointer to setup documentation" goal of this PR.
Suggestion: Please verify this path resolves to a live page in the rendered site build (the PR notes a docs link was verified, but confirming the exact slug matches the deployed route avoids a 404).
| Bidi Agents connect with frontier voice models like GPT-Realtime-2.1, Gemini 3.8 Live, and the [newly released Amazon Nova 2.5 Sonic](https://aws.amazon.com/about-aws/whats-new/2026/10/amazon-nova-2.5-Sonic/). The improved reasoning from Nova 2.5 Sonic shows a leap for what's possible with voice agents. We recently prototyped a Bidi Agent powered by this model for Zoom calls, so it could chat with us during brainstorms and run web searches to fill in missing information. Bidi Agents make it easy to switch to a different provider while remaining observable. Our telemetry captures OTEL spans for each session, model response, and tool call, including reported token usage. | ||
|
|
||
| Bidi Agents also ensure the conversation can continue for as long as you want it to. Most voice models currently enforce session limits that could interrupt a conversation, however, Bidi Agents keeps the connection alive by restarting it in the background before the session cap is hit. For example, Nova 2.5 Sonic's 8 minute limit can be configured in Bidi Agents to run as long as you want it to. | ||
| Bidi Agents also ensure the conversation can continue for as long as you want it to. Most voice models currently enforce session limits that could interrupt a conversation, but Bidi Agents keep the connection alive by restarting it in the background before the session cap is hit. For example, Nova 2.5 Sonic's 8-minute limit can be configured in Bidi Agents to run as long as you want it to. |
There was a problem hiding this comment.
Suggestion
Issue: The model names GPT-Realtime-2.1 and Gemini 3.8 Live are stated here as frontier voice models; the PR corrects the OpenAI name but Gemini is unchanged.
Suggestion: Please verify both product names (and the Nova 2.5 Sonic reference) match the official vendor branding at publish time, since customer-facing announcements are sensitive to naming accuracy.
| Bidi Agents connect with frontier voice models like GPT-Realtime-2.1, Gemini 3.8 Live, and the [newly released Amazon Nova 2.5 Sonic](https://aws.amazon.com/about-aws/whats-new/2026/10/amazon-nova-2.5-Sonic/). The improved reasoning from Nova 2.5 Sonic shows a leap for what's possible with voice agents. We recently prototyped a Bidi Agent powered by this model for Zoom calls, so it could chat with us during brainstorms and run web searches to fill in missing information. Bidi Agents make it easy to switch to a different provider while remaining observable. Our telemetry captures OTEL spans for each session, model response, and tool call, including reported token usage. | ||
|
|
||
| Bidi Agents also ensure the conversation can continue for as long as you want it to. Most voice models currently enforce session limits that could interrupt a conversation, however, Bidi Agents keeps the connection alive by restarting it in the background before the session cap is hit. For example, Nova 2.5 Sonic's 8 minute limit can be configured in Bidi Agents to run as long as you want it to. | ||
| Bidi Agents also ensure the conversation can continue for as long as you want it to. Most voice models currently enforce session limits that could interrupt a conversation, but Bidi Agents keep the connection alive by restarting it in the background before the session cap is hit. For example, Nova 2.5 Sonic's 8-minute limit can be configured in Bidi Agents to run as long as you want it to. |
There was a problem hiding this comment.
Suggestion
Issue: "Our telemetry captures OTEL spans for each session, model response, and tool call, including reported token usage" — the rewrite drops the explicit "token usage" span that the original listed as its own item, folding it into "reported token usage."
Suggestion: Please confirm the new phrasing still accurately reflects what the telemetry emits (i.e., whether token usage is a separate span or an attribute), so the doc doesn't overstate or understate observability behavior.
GitHub Actions
Failure
Documentation Preview FailedThe documentation deployment encountered an error. Please check the deployment logs for more details. |
Replica of
https://github.com/strands-agents/harness-sdk/pull/4904(URL in backticks to avoid a cross-reference on the original PR) for reviewer evaluation. Branches mirror the original PR's base and head commits.The review below was produced by the optimized Strandly reviewer (variant
v3-reason, optimization runaeb3ee, agent model Opus 4.8) during validation on PRs it never saw in training. Compare with the human and Strandly reviews on the original PR.