docs(v4): clarify evaluator transition bridge

[marliessophie]

Summary

• document the temporary trace input/output bridge for existing trace-level evaluators
• name the deprecated Python and JS/TS methods users may need during the transition
• clarify that future multi-span support will use the observation-centric model
• replace the generic set_trace_io wording in the v4 comparison table

Stacking

• based on Max's v4 documentation PR docs: v4 release on cloud #3330
• complements the detailed evaluator upgrade guide in docs(evals): update trace-level evaluator upgrade guide #3323

Verification

• pnpm exec prettier --experimental-cli --check components-mdx/v4/step-evaluators.mdx content/docs/v4.mdx
• node scripts/check-h1-headings.js
• git diff --check

Full build and link checks were not run for this two-line documentation follow-up.

Greptile Summary

This PR clarifies the evaluator transition bridge in the v4 documentation. The main changes are:

• Names the deprecated Python and JavaScript/TypeScript trace I/O methods.
• Explains when to retain those methods during evaluator migration.
• Reaffirms the observation-centric model for future multi-span support.
• Updates the v4 comparison table with the temporary bridge wording.

Confidence Score: 5/5

This looks safe to merge.

• No blocking issues found in the changed documentation.
• The method names and deprecation guidance match the related SDK upgrade and evaluator migration documentation.

_{Reviews (1): Last reviewed commit: "docs(v4): clarify evaluator transition b..." | Re-trigger Greptile}

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
langfuse-docs	Ready	Preview, Comment	Jul 22, 2026 10:02am

[github-actions]

@claude review

[claude]

LGTM — small, accurate doc clarification; the one nit flagged inline (OTel/direct-API readers not covered by the new bridge wording) doesn't block merge.

What was reviewed: the updated bridge sentence in step-evaluators.mdx and the corresponding comparison-table row in v4.mdx; confirmed the method names (set_current_trace_io/span.set_trace_io, setActiveTraceIO/span.setTraceIO) match existing deprecated APIs; checked for a md-override counterpart to v4.mdx (none exists, so no sync needed).

Extended reasoning...

Overview

Two-line documentation change: rewords the evaluator-transition guidance in the shared components-mdx/v4/step-evaluators.mdx snippet (rendered across all four v4 upgrade paths) to describe a temporary bridge using deprecated trace I/O methods, and updates one row of the before/after comparison table in content/docs/v4.mdx to match. No code, config, or build logic is touched.

Security risks

None. This is prose-only documentation content; no user input, auth, or executable code paths are affected.

Level of scrutiny

Low. This is a narrow, self-contained doc wording fix stacked on an already-merged v4 docs PR. The only substantive concern — that the new parenthetical naming Python/JS SDK methods doesn't give OTel/direct-API readers an actionable equivalent — was already caught and will be posted as an inline nit. It doesn't misinform that audience (the clause is self-scoped to Python/JS) and doesn't regress prior behavior, so it doesn't warrant blocking.

Other factors

No corresponding md-override file exists for content/docs/v4.mdx, so there's no sync requirement. The PR description notes local prettier/H1 checks were run; given the tiny, text-only diff, that's sufficient without a full build.

[claude]

🟡 The new bridge sentence in step-evaluators.mdx only mentions Python (set_current_trace_io()/span.set_trace_io()) and JS/TS (setActiveTraceIO()/span.setTraceIO()) methods for keeping trace I/O populated during the evaluator transition, but this shared component is also rendered in the "Direct API and OpenTelemetry users" section, whose readers by definition aren't using either SDK. Consider scoping the parenthetical (e.g. "if you use the Python/JS SDK") or adding a short note for OTel/direct-API users so that audience isn't left assuming an inapplicable method.

Extended reasoning...

components-mdx/v4/step-evaluators.mdx is rendered via <V4StepEvaluators /> in all four upgrade paths on content/docs/v4.mdx, including ### Direct API and OpenTelemetry users [#api-otel] (line ~181). That section's own Ingestion step is explicit that its audience is not using the Langfuse Python/JS SDK — it instructs readers to point a generic OTel span exporter at POST /api/public/otel/v1/traces, with SDK users routed to the separate cloud-older-sdks / cloud-current-sdks sections instead.

The sentence this PR adds says: "If an existing evaluator reads trace input or output, keep supplying those values with the deprecated trace I/O methods (set_current_trace_io() or span.set_trace_io() in Python; setActiveTraceIO() or span.setTraceIO() in JS/TS) until the new evaluator is validated." Walking through the api-otel reader's experience: they land on this Step expecting guidance for their setup, see two method names that are explicitly qualified "in Python"/"in JS/TS", and have no way to apply either one, since they aren't calling into an SDK at all. There's no OTel/API-level equivalent offered (e.g., an attribute name to set on a span via the raw ingestion payload), so this subset of readers gets no actionable instruction for the trace-I/O bridge if they have an existing trace-level evaluator they want to keep working during the transition.

That said, the risk here is genuinely limited, which is why this should not block merge. Before this PR, the shared sentence was just "Migrate trace-level LLM-as-a-judge evaluators to the observation level" — equally generic for all four audiences, with the same link to the migration FAQ. This PR only adds an optional bridging clause for SDK users; it doesn't remove or weaken anything OTel users previously had, and the core instruction (upgrade to observation-level evaluators, see the migration guide) still applies to and works for every audience, including api-otel. The added clause is clearly self-scoped ("in Python", "in JS/TS"), so an OTel reader will recognize it doesn't apply to them rather than being misled into calling a nonexistent method. It's also not obviously fixable within this doc sentence alone — if OTel/direct-API ingestion has no equivalent way to keep trace I/O populated (this looks like a genuine product gap, not a doc omission), the best available fix is to explicitly scope the parenthetical to SDK users, so OTel readers aren't left wondering why it doesn't apply to them, or to add a short pointer to whatever mechanism (if any) exists for raw OTel spans.

Net effect: a documentation completeness gap affecting one of four audiences for an edge case (readers with a pre-existing trace-level evaluator, mid-transition), not an incorrect or misleading instruction. Recommend scoping the parenthetical (e.g. "if you use the Python or JS/TS SDK, keep supplying...") or adding a one-line OTel-specific caveat, but this is a nit rather than a blocking issue.