feat(opencode): show token throughput metrics - #12434
Merged
marius-kilocode merged 19 commits intoJul 23, 2026
Merged
Conversation
Githubguy132010
marked this pull request as ready for review
July 21, 2026 17:33
![kilo-code-bot[bot]](https://avatars.githubusercontent.com/in/2193792?s=60&v=4){kind=small}
Contributor
Code Review SummaryStatus: No Issues Found | Recommendation: Merge Reviewed the incremental diff since the last review (commit
No new bugs, style violations, or fork-hygiene issues introduced by this commit. All 40 existing inline comments on this PR remain stale ( Files Reviewed (1 file, incremental)
Previous Review Summaries (8 snapshots, latest commit 582cf4d)Current summary above is authoritative. Previous snapshots are kept for context only. Previous review (commit 582cf4d)Status: No Issues Found | Recommendation: Merge Reviewed the incremental diff since the last review (verified via
All findings from the previous review round (guard blocking top-level settings persistence, Files Reviewed (28 files, incremental)
Previous review (commit 65e552c)Status: No Issues Found | Recommendation: Merge All findings from the previous review have been addressed in the latest commits:
Reviewed the incremental diff (34 files, ~495/-163 lines) covering the throughput wire-shape addition ( Files Reviewed (34 files)
Previous review (commit 41270b0)Status: 6 Issues Found | Recommendation: Address before merge Overview
Note: No new commits with file changes since the last review. The only new commit ( Issue Details (click to expand)CRITICAL
WARNING
SUGGESTION
Files Reviewed (0 new files — no diff since last review)No file changes between the previously reviewed commit and current Fix these issues in Kilo Cloud Previous review (commit 405a0c1)Status: 6 Issues Found | Recommendation: Address before merge Overview
Note: No new commits with file changes since the last review. The only new commit ( Issue Details (click to expand)CRITICAL
WARNING
SUGGESTION
Files Reviewed (0 new files — no diff since last review)No file changes between the previously reviewed commit and current Fix these issues in Kilo Cloud Previous review (commit aa3fe36)Status: 6 Issues Found | Recommendation: Address before merge Overview
Note: The branch was rewritten again since the last review (the previous review commit is no longer in history), so this is a fresh full review. Good news: the merge-commit regression that had reverted several fixes (root Issue Details (click to expand)CRITICAL
WARNING
SUGGESTION
Files Reviewed (46 files)
Fix these issues in Kilo Cloud Previous review (commit ec91f2a)Status: 9 Issues Found | Recommendation: Address before merge Overview
Note: The branch history was rewritten since the last review (previous review commit is no longer an ancestor of HEAD), so this pass falls back to a full review. All issues from the prior review remain unresolved — the merge that landed on this branch ( Issue Details (click to expand)CRITICAL
WARNING
SUGGESTION
Files Reviewed (48 files)
Fix these issues in Kilo Cloud Previous review (commit aeff95a)Status: No Issues Found | Recommendation: Merge All issues from the previous review have been addressed in this update:
The remaining previously-flagged item ( Files Reviewed in this incremental pass (28 files)
Previous review (commit 040dde6)Status: 7 Issues Found | Recommendation: Address before merge Overview
Issue Details (click to expand)CRITICAL
WARNING
SUGGESTION
Files Reviewed (53 files)
[Snapshot truncated.] Additional previous summary content was truncated to keep this comment within platform limits. Reviewed by claude-sonnet-5 · Input: 44 · Output: 10.7K · Cached: 1.3M Review guidance: REVIEW.md from base branch |
Githubguy132010
force-pushed
the
feat/token-throughput-v2
branch
from
July 22, 2026 08:46
4c4f567 to
ec91f2a
Compare
Githubguy132010
force-pushed
the
feat/token-throughput-v2
branch
from
July 22, 2026 09:09
025b2ea to
ec91f2a
Compare
Githubguy132010
force-pushed
the
feat/token-throughput-v2
branch
from
July 22, 2026 10:13
cb1c1d1 to
89ff124
Compare
Capture prompt-processing and text-generation tokens/sec on every StepFinishPart. The metrics helper prefers provider-reported rates from llama.cpp / vLLM timings and falls back to wall-clock computation. A new kilocode tui usage route renders PP/TG inline. chore(sdk): regenerate types for StepFinishPart.metrics feat(tui): render PP/TG in sidebar usage panel feat(vscode): per-message and aggregated token throughput display Surface throughput on each AssistantMessage badge (behind the showTokenThroughput toggle) and as a compact PP/TG row in the expanded TaskHeader. Adds session helpers, i18n entries in 20 locales, and StepFinishPart.metrics to extension/webview messages. Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com>
…er branch - Replace the dead sendThroughputSetting() private with the shared buildThroughputSettingMessage() helper and add validThroughputSetting to handleUpdateSetting so the showTokenThroughput setting has the same guard as the chat/indexing twins (fixes Knip regression). - Bind the DisplayTab Switch to the local settings draft so the toggle flips on click instead of waiting for a Save round-trip (the user-facing kill switch for #6579). - Narrow StepThroughputMetrics.source to "computed"; backend hard-codes computed metrics today because the upstream AI SDK drops provider timings. Drop the unused provider branches from AssistantMessage and TaskHeader so the rendering code has no dead paths. Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com>
- Replace inline PP/TG labels in the CLI sidebar with a throughputLabel constant in model-usage so a future i18n sweep is one file instead of every rendering site. - Tighten isStepMetrics in sidebar-usage back to a real discriminator check after dropping the unreachable "provider" union member. - Drop formatPP/formatTG exports from model-usage since callers already use the shared formatRateValue; mirror the swap in the TUI usage test. Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com>
…nd test computed-only - aggregateMetrics adopts the first non-empty computed sample per field across every step-finish in the session, replacing the dead provider- ranked last-wins strategy that shipped with the unreachable branch. - Share the throughputVisible signal through DisplayProvider so every AssistantMessage and the TaskHeader row react to a single onMount requestThroughputSetting round-trip, instead of each message posting its own handshake. - Drop the unused routes/session/usage.tsx TUI route (no remaining imports) and add the chat-layout badge/header pill styles it was gating on. - Refresh session-utils tests to exercise only source: "computed" samples and follow the new first-wins rule. Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com>
…der branch Seed the DisplayTab "Show Token Throughput" Switch on initial load by mirroring the throughputSettingLoaded message into settings() (same pattern as chat.shiftTabCyclesVariant). Without this, a persisted-true setting renders unchecked on first open because the Switch was bound to settings()["showTokenThroughput"] but no handler ever populated it. Drop the dead data-source attributes on the per-message and task-header throughput surfaces now that StepThroughputMetrics.source is narrowed to "computed" only — the attribute was always the literal string. Drop the unreachable chat.throughput.badge.provider and .chat.throughput.badge.tooltip.provider i18n keys across all 20 locales. The badges no longer branch on source === "provider" since the provider-source branch is removed (the AI SDK adapter upstream strips llama.cpp timings before they reach providerMetadata). Co-Authored-By: Claude <noreply@anthropic.com> Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com>
PP (prompt-processing rate) has no signal in this build: the AI SDK adapter upstream strips llama.cpp's `prompt_per_second` before it reaches providerMetadata, and computeMetrics has nothing else to derive it from. Ship the TG (text-generation) rate only — the UI no longer renders the "PP –" placeholder that made the feature look broken. CLI sidebar drops the PP row; per-message badge and aggregated header pill both lose the "PP – ·" prefix. The wire shape keeps the optional prompt field so the follow-up that wires the upstream metadataExtractor can populate it without another schema bump. The `throughputLabel` constant on the opencode side and the `formatPP` helper on the webview side are removed; tests that fabricated prompt values are pruned to match. Co-Authored-By: Claude <noreply@anthropic.com> Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com>
The standalone blue pills read as loud for what is secondary session info. Move the aggregated TG into the existing Tokens row as another spanned value (alongside ↑ input, ↑ cache, ↓ output) and restyle the per-message badge as plain text in descriptionForeground so both surfaces match the tokens family. TaskUsage now accepts a `throughput` prop and renders `TG <rate> t/s` inline in the Summary component when the toggle is on. TaskHeader no longer emits a standalone [data-slot="task-header-throughput"] element; its [data-slot="task-header-throughput"] CSS rule is removed. The throughputText / throughputTooltip memos and the unused formatTG import are dropped — the values flow straight into TaskUsage. Co-Authored-By: Claude <noreply@anthropic.com> Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com>
…Message The "Merge origin/main into feat/token-throughput-v2" resolution kept the throughput branch's memory-badge code (already removed from main by 28d015f), which broke the kilo-ui-contract test and the i18n-keys test. Drop the dead code: `useMemory`/`MemoryMarkerMeta` imports, `mem`, the `meta`/`recall`/`fmt`/`count`/`items`/`verbose` createMemos, the `tip` function, and the `<Show when={mem.enabled() && recall()}>` block. The file lands at 349 lines (down from 391), matching main + throughput only. Verified locally: - i18n-keys + kilo-ui-contract: 53 pass, 0 fail - Full kilo-vscode suite: failures 138 → 136 (+2 from the two fixes) Co-Authored-By: Claude <noreply@anthropic.com> Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com>
…Message Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com>
Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com>
- Rename 'TG' to 'Generation speed' in en.ts and add a 'gauge' icon to packages/kilo-ui so the per-message badge and the Tokens row show '<icon> Generation speed <rate> t/s' instead of the cryptic 'TG <rate>'. Centralize the opencode sidebar label in throughputLabel.generation. - Switch aggregateMetrics (both webview and CLI) to the latest non-empty step-finish snapshot so only the most recent assistant turn's generation rate is shown rather than a session-wide aggregate. Update tests and comments to match. - Translate the throughput strings in no.ts to Norwegian; mirror the new key shape across the other locales (English fallback for untranslated strings). Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com>
Githubguy132010
force-pushed
the
feat/token-throughput-v2
branch
from
July 22, 2026 10:16
89ff124 to
a310090
Compare
… formatting The throughput rebases left a duplicated .vscode-session-turn-diffs selector and let three files drift from prettier's expectations. Fix the CSS unclosed-block (which broke the Storybook preview build) and re-run prettier --write on the touched files. Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com>
Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com>
Kilo:
- CRITICAL: Remove the section === '' guard in KiloProvider.ts that
blocked persistence of every top-level setting key. The throughput
validator is now redundant, so drop it from throughput-settings.ts.
- WARNING: Reset samples in sidebar-usage.tsx when props.session_id
changes via a keyed createEffect, so a session switch no longer
blends step-finish metrics from the previous session and the array
no longer grows without bound across long-lived plugin instances.
- WARNING: Pass {speed} to language.t for the TaskUsage throughput
tooltip and reuse the shared formatTG helper instead of reformatting
the value inline. Drop the dead [data-component='assistant-memory-badge']
rule whose target component no longer exists in the tree.
- SUGGESTION: Drop redundant guards in model-usage.ts (undefined check
after Number.isFinite, and the ?? 0 on an already-required number
field). Use typeof === 'number' for the type narrowing.
Address Marius's review of the throughput UI:
Calculation
- Persist per-step timing (start/end/elapsed) on step-start and
step-finish parts in the session processor.
- Add wire schemas in core/src/v1/session.ts and packages/sdk/openapi.json
so the new time field round-trips end-to-end.
- Replace the last-wins 'latest step rate' snapshot with a weighted
aggregate: sum(output + reasoning tokens) / sum(active generation
duration) across the turn's step-finish parts. Tool execution and
idle waiting are excluded.
- The CLI sidebar (model-usage.ts) gains the same weighted semantics
when timing is available, falling back to last-wins otherwise so
older callers keep working.
Presentation
- Strip the per-message badge to plain muted text (no icon, no label,
no border). The chip in the upstream action row reads as metadata.
- Move throughput out of the task header Tokens row so each turn owns
its own value (no flicker across turns, single source of truth).
- Read the throughput memo from the full message parts in the data
store rather than the chunked row slice, so step-finish in any
chunk produces the badge.
i18n
- Replace chat.throughput.speed.{label,row,tooltip,tooltip.missing}
with chat.throughput.tooltip and chat.throughput.tooltip.missing
across all 19 locale files.
Tests
- Add messageThroughput and sessionThroughput describe blocks
exercising the weighted aggregate across multiple steps.
- Cover weighted + fallback paths in the CLI aggregateMetrics tests.
Collaborator
I propose we move it up here instead.
|
Move the throughput badge from a footer line below the assistant message into the copy/feedback action row of the text part that carries the copy button. This avoids the extra vertical space the footer consumed. Also apply prettier formatting to drifted PR files (i18n line wraps, TaskHeader/session-utils/test reflows).
marius-kilocode
approved these changes
Jul 23, 2026
marius-kilocode
left a comment
marius-kilocode
left a comment
Collaborator
There was a problem hiding this comment.
Reviewed the full diff after the latest commits: backend schema additions are additive and optional (legacy sessions keep decoding), the processor timing/metrics plumbing is guarded against missing step-start, webview surfaces (per-message badge now inline beside the copy/feedback row + task header) share one toggle, and tests cover the aggregation paths. Fixed the changeset package name (@kilocode/kilo -> @kilocode/cli) and the missing kilocode_change marker on the step-start time field. No blockers.
marius-kilocode
enabled auto-merge (squash)
July 23, 2026 09:15
Contributor
Author
Alright, I would like to thank you for collaborating with me on the PR. It really helps me move faster. |
t7tran
pushed a commit
to t7tran/kilocode
that referenced
this pull request
Aug 14, 2026
* feat(opencode): record per-step token throughput metrics Capture prompt-processing and text-generation tokens/sec on every StepFinishPart. The metrics helper prefers provider-reported rates from llama.cpp / vLLM timings and falls back to wall-clock computation. A new kilocode tui usage route renders PP/TG inline. chore(sdk): regenerate types for StepFinishPart.metrics feat(tui): render PP/TG in sidebar usage panel feat(vscode): per-message and aggregated token throughput display Surface throughput on each AssistantMessage badge (behind the showTokenThroughput toggle) and as a compact PP/TG row in the expanded TaskHeader. Adds session helpers, i18n entries in 20 locales, and StepFinishPart.metrics to extension/webview messages. Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com> * fix(vscode): wire token throughput toggle and drop unreachable provider branch - Replace the dead sendThroughputSetting() private with the shared buildThroughputSettingMessage() helper and add validThroughputSetting to handleUpdateSetting so the showTokenThroughput setting has the same guard as the chat/indexing twins (fixes Knip regression). - Bind the DisplayTab Switch to the local settings draft so the toggle flips on click instead of waiting for a Save round-trip (the user-facing kill switch for Kilo-Org#6579). - Narrow StepThroughputMetrics.source to "computed"; backend hard-codes computed metrics today because the upstream AI SDK drops provider timings. Drop the unused provider branches from AssistantMessage and TaskHeader so the rendering code has no dead paths. Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com> * fix(opencode): centralize token throughput labels and tighten type guard - Replace inline PP/TG labels in the CLI sidebar with a throughputLabel constant in model-usage so a future i18n sweep is one file instead of every rendering site. - Tighten isStepMetrics in sidebar-usage back to a real discriminator check after dropping the unreachable "provider" union member. - Drop formatPP/formatTG exports from model-usage since callers already use the shared formatRateValue; mirror the swap in the TUI usage test. Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com> * feat(token-throughput-v2): wire aggregation through DisplayProvider and test computed-only - aggregateMetrics adopts the first non-empty computed sample per field across every step-finish in the session, replacing the dead provider- ranked last-wins strategy that shipped with the unreachable branch. - Share the throughputVisible signal through DisplayProvider so every AssistantMessage and the TaskHeader row react to a single onMount requestThroughputSetting round-trip, instead of each message posting its own handshake. - Drop the unused routes/session/usage.tsx TUI route (no remaining imports) and add the chat-layout badge/header pill styles it was gating on. - Refresh session-utils tests to exercise only source: "computed" samples and follow the new first-wins rule. Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com> * fix(token-throughput-v2): seed settings on hydration, drop dead provider branch Seed the DisplayTab "Show Token Throughput" Switch on initial load by mirroring the throughputSettingLoaded message into settings() (same pattern as chat.shiftTabCyclesVariant). Without this, a persisted-true setting renders unchecked on first open because the Switch was bound to settings()["showTokenThroughput"] but no handler ever populated it. Drop the dead data-source attributes on the per-message and task-header throughput surfaces now that StepThroughputMetrics.source is narrowed to "computed" only — the attribute was always the literal string. Drop the unreachable chat.throughput.badge.provider and .chat.throughput.badge.tooltip.provider i18n keys across all 20 locales. The badges no longer branch on source === "provider" since the provider-source branch is removed (the AI SDK adapter upstream strips llama.cpp timings before they reach providerMetadata). Co-Authored-By: Claude <noreply@anthropic.com> Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com> * fix(token-throughput-v2): drop PP display until llama.cpp wiring lands PP (prompt-processing rate) has no signal in this build: the AI SDK adapter upstream strips llama.cpp's `prompt_per_second` before it reaches providerMetadata, and computeMetrics has nothing else to derive it from. Ship the TG (text-generation) rate only — the UI no longer renders the "PP –" placeholder that made the feature look broken. CLI sidebar drops the PP row; per-message badge and aggregated header pill both lose the "PP – ·" prefix. The wire shape keeps the optional prompt field so the follow-up that wires the upstream metadataExtractor can populate it without another schema bump. The `throughputLabel` constant on the opencode side and the `formatPP` helper on the webview side are removed; tests that fabricated prompt values are pruned to match. Co-Authored-By: Claude <noreply@anthropic.com> Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com> * fix(token-throughput-v2): integrate TG into Tokens row, plain-text style The standalone blue pills read as loud for what is secondary session info. Move the aggregated TG into the existing Tokens row as another spanned value (alongside ↑ input, ↑ cache, ↓ output) and restyle the per-message badge as plain text in descriptionForeground so both surfaces match the tokens family. TaskUsage now accepts a `throughput` prop and renders `TG <rate> t/s` inline in the Summary component when the toggle is on. TaskHeader no longer emits a standalone [data-slot="task-header-throughput"] element; its [data-slot="task-header-throughput"] CSS rule is removed. The throughputText / throughputTooltip memos and the unused formatTG import are dropped — the values flow straight into TaskUsage. Co-Authored-By: Claude <noreply@anthropic.com> Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com> * fix(token-throughput-v2): finish removing memory badge from AssistantMessage The "Merge origin/main into feat/token-throughput-v2" resolution kept the throughput branch's memory-badge code (already removed from main by 8ec35e3), which broke the kilo-ui-contract test and the i18n-keys test. Drop the dead code: `useMemory`/`MemoryMarkerMeta` imports, `mem`, the `meta`/`recall`/`fmt`/`count`/`items`/`verbose` createMemos, the `tip` function, and the `<Show when={mem.enabled() && recall()}>` block. The file lands at 349 lines (down from 391), matching main + throughput only. Verified locally: - i18n-keys + kilo-ui-contract: 53 pass, 0 fail - Full kilo-vscode suite: failures 138 → 136 (+2 from the two fixes) Co-Authored-By: Claude <noreply@anthropic.com> Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com> * fix(token-throughput-v2): finish removing memory badge from AssistantMessage Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com> * formatting fixes Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com> * fix(token-throughput-v2): address Marius review comments - Rename 'TG' to 'Generation speed' in en.ts and add a 'gauge' icon to packages/kilo-ui so the per-message badge and the Tokens row show '<icon> Generation speed <rate> t/s' instead of the cryptic 'TG <rate>'. Centralize the opencode sidebar label in throughputLabel.generation. - Switch aggregateMetrics (both webview and CLI) to the latest non-empty step-finish snapshot so only the most recent assistant turn's generation rate is shown rather than a session-wide aggregate. Update tests and comments to match. - Translate the throughput strings in no.ts to Norwegian; mirror the new key shape across the other locales (English fallback for untranslated strings). Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com> * fix(token-throughput-v2): close unclosed CSS block and apply prettier formatting The throughput rebases left a duplicated .vscode-session-turn-diffs selector and let three files drift from prettier's expectations. Fix the CSS unclosed-block (which broke the Storybook preview build) and re-run prettier --write on the touched files. Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com> * ci: re-run after fixing CSS unclosed-block + prettier drift Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com> * fix(token-throughput-v2): address kilo-code-bot review Kilo: - CRITICAL: Remove the section === '' guard in KiloProvider.ts that blocked persistence of every top-level setting key. The throughput validator is now redundant, so drop it from throughput-settings.ts. - WARNING: Reset samples in sidebar-usage.tsx when props.session_id changes via a keyed createEffect, so a session switch no longer blends step-finish metrics from the previous session and the array no longer grows without bound across long-lived plugin instances. - WARNING: Pass {speed} to language.t for the TaskUsage throughput tooltip and reuse the shared formatTG helper instead of reformatting the value inline. Drop the dead [data-component='assistant-memory-badge'] rule whose target component no longer exists in the tree. - SUGGESTION: Drop redundant guards in model-usage.ts (undefined check after Number.isFinite, and the ?? 0 on an already-required number field). Use typeof === 'number' for the type narrowing. * feat(token-throughput-v2): weighted per-turn rate, plain text footer Address Marius's review of the throughput UI: Calculation - Persist per-step timing (start/end/elapsed) on step-start and step-finish parts in the session processor. - Add wire schemas in core/src/v1/session.ts and packages/sdk/openapi.json so the new time field round-trips end-to-end. - Replace the last-wins 'latest step rate' snapshot with a weighted aggregate: sum(output + reasoning tokens) / sum(active generation duration) across the turn's step-finish parts. Tool execution and idle waiting are excluded. - The CLI sidebar (model-usage.ts) gains the same weighted semantics when timing is available, falling back to last-wins otherwise so older callers keep working. Presentation - Strip the per-message badge to plain muted text (no icon, no label, no border). The chip in the upstream action row reads as metadata. - Move throughput out of the task header Tokens row so each turn owns its own value (no flicker across turns, single source of truth). - Read the throughput memo from the full message parts in the data store rather than the chunked row slice, so step-finish in any chunk produces the badge. i18n - Replace chat.throughput.speed.{label,row,tooltip,tooltip.missing} with chat.throughput.tooltip and chat.throughput.tooltip.missing across all 19 locale files. Tests - Add messageThroughput and sessionThroughput describe blocks exercising the weighted aggregate across multiple steps. - Cover weighted + fallback paths in the CLI aggregateMetrics tests. * fix(token-throughput-v2): render t/s inline beside copy/feedback buttons Move the throughput badge from a footer line below the assistant message into the copy/feedback action row of the text part that carries the copy button. This avoids the extra vertical space the footer consumed. Also apply prettier formatting to drifted PR files (i18n line wraps, TaskHeader/session-utils/test reflows). * fix(token-throughput-v2): correct changeset package name to @kilocode/cli * fix(token-throughput-v2): annotate step-start time field with kilocode_change --------- Co-authored-by: Thomas Brugman <thomas@kilocode.ai> Co-authored-by: kiloconnect[bot] <240665456+kiloconnect[bot]@users.noreply.github.com> Co-authored-by: marius-kilocode <marius@kilocode.ai>
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
2 participants
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.
Issue
Closes #6579
Context
Adds tokens-per-second throughput metrics for #6579. The backend records a per-step generation rate on
StepFinishPart; the Kilo CLI sidebar surfaces aTGrow in its usage panel, and the VSCode webview integrates TG into the existing Tokens row at the top of the task header plus a per-message line beneath each assistant response. Both VSCode surfaces use the same plain-text styling as the surrounding tokens so throughput reads as one family of secondary info rather than a separate pill. A new "Show Token Throughput" toggle in Display settings (off by default) controls both VSCode surfaces.Replaces the approach from closed PR #9773 — that one was shelved because Marius wanted a clearer UX before going forward and the token indicator stopped showing after upstream merges. This iteration narrows the wire shape, gates display behind a toggle, integrates TG into existing UI rather than adding standalone elements, and runs every change through three rounds of adversarial review (correctness / UX / merge-conflict) before shipping.
Implementation
Backend (
packages/opencode/src/kilocode/session/metrics.ts): newcomputeMetricshelper derives generation rate from(output + reasoning) tokens * 1000 / elapsedMs. Optionalmetricsfield added toStepFinishPartschema inpackages/core/src/v1/session.tswithkilocode_changemarkers; SDK regenerated to expose the new field.CLI (
packages/opencode/src/kilocode/plugins/sidebar-usage.tsx): singleTGrow in the usage panel, hidden when no metrics are available. Local aggregation via token-weighted mean across step-finish parts.VSCode (
packages/kilo-vscode/webview-ui/):TaskUsage.tsx(alongside ↑ input, ↑ cache, ↓ output) — no separate element.TaskUsageaccepts athroughput?: numberprop and emitsTG <rate> t/sas a plain-text span when present, matching the surroundingtask-header-tokens-valuestyle.AssistantMessage.tsxrenders as plain text usingvar(--vscode-descriptionForeground)andfont-variant-numeric: tabular-nums— same tone as the Tokens row, no pill background or border.display.throughputVisible()signal hoisted toDisplayProviderso toggling once updates both without round-trips.settings()via the existingchatSettingsLoadedpattern;onDidChangeConfigurationwatcher pushes runtime setting changes to the webview.PP deferred: the AI SDK adapter at
packages/opencode/src/session/llm/ai-sdk.ts:64wraps raw usage as{ aiSdk: item.raw }, but@ai-sdk/openai-compatible'sconvertOpenAICompatibleChatUsagereturnsraw: usage— the response-leveltimingsobject (where llama.cpp putsprompt_per_second/predicted_per_second) never reachesproviderMetadata. PP cannot be derived without patching the sharedcreateOpenAICompatiblecall atpackages/opencode/src/provider/provider.ts:131(out of scope for a Kilo-only change). Follow-up tracked separately; the wire shape keeps the optionalpromptfield so the schema doesn't need to change.Provider source dropped: same root cause. The
providerRatebranch was opportunistic detection of provider-reported rates — but with the upstream blocker, it's dead code.computeMetricsalways emitssource: "computed"; consumers (CLI sidebar, webview surfaces) handle the single literal.Screenshots / Video
Schermopname.2026-07-21.om.19.25.04.mov
How to Test
Manual/local verification
cd packages/kilo-vscode && bun run extension. Toggle "Show Token Throughput" in Display settings. Run a non-trivial assistant prompt. Confirm the TG value appears inline in the Tokens row at the top of the task header and as plain text under each assistant message. Toggle off — both surfaces should hide immediately without reload. Editsettings.jsondirectly to flip the value — same behavior.cd packages/opencode && bun run build && ./bin/kilo. Open the TUI sidebar, run a session, confirm theTGrow appears in the usage panel alongside Input/Output/Reasoning/Cache.computeMetricsreturns undefined whengenerated <= 0).Reviewer test steps
cd packages/kilo-vscode && bun run extensionto launch the dev VS Code window.TG <rate> t/sappears as plain text under the assistant message.TG <rate> t/sappears inline there too (after the ↓ output value), in the same plain-text style as the surrounding tokens.~/.config/Code/User/settings.jsonto flipkilo-code.new.showTokenThroughput. Confirm the toggle reflects without reload.cd packages/opencode && bun run build && ./bin/kilo. Run a session, confirmTGrow appears in the sidebar usage panel.Blocked checks and substitute verification
bun turbo typecheckwas excluded by design — it fans out to a Gradle/Kotlin build ofpackages/kilo-jetbrains(~9 min cold cache) that is unrelated to this feature. Per-package checks (packages/opencodetypecheck + targeted tests;packages/kilo-vscodetypecheck + lint + test:unit + knip + check-kilocode-change) cover the touched code and were green.bun test:unitinpackages/kilo-vscode/reports 3251 pass / 1 fail. The single failing test istests/unit/worktree-manager.test.ts:986(WorktreeManager.resolveStartPoint > returns bare branch + remote when remote exists), which is a pre-existing failure onmainunrelated to this branch (the test file is untouched here).Verification
The diff went through three rounds of adversarial review (correctness, UX, merge-conflict lenses), each with refuter-style finding verification. UI was iterated inline after review: pill styling replaced with plain text matched to the Tokens row, aggregated TG moved into the Tokens row itself rather than a standalone element. Final state: aggregated TG inline in
TaskUsage, per-message TG as a plain-text line, both gated bydisplay.throughputVisible(), toggle hydrates fromsettings()and reflects runtime changes viaonDidChangeConfiguration.Checklist
.changeset/token-throughput-v2.md,minorbump on@kilocode/kiloand@kilocode/sdk)Get in Touch
Discord: thomas07374