feat(usage): include Grok Build in the usage view

[EliasofIX]

Summary

The usage page only scanned Claude Code and Codex session transcripts, so Grok Build work never showed up even though sessions under ~/.grok already carry per-turn token and cost data.

This adds Grok as a third usage provider end to end:

• Parse turn_completed usage from Grok updates.jsonl (per-model modelUsage, cache breakdown, costUsdTicks at 10^10 ticks per USD)
• Scan $GROK_HOME/sessions (default ~/.grok/sessions), restricted to updates.jsonl so large chat/events logs stay out of the cold path
• Surface Grok Build in the web and mobile usage charts/labels
• Bump the usage contract version for the new provider kind

Test plan

• Open Usage after pairing a local server that has Grok sessions in ~/.grok
• Confirm Grok Build appears in the provider split, daily chart, and model table
• Confirm Claude/Codex totals still match the previous behavior
• vp test run apps/server/src/usage/usageTranscripts.test.ts

Made with Grok Build (xAI).

---

Note

Medium Risk
New transcript parsing and cost pro-rating logic could mis-attribute tokens or dollars for edge-case Grok logs, but changes are confined to usage scanning/aggregation with broad test coverage and a contract version bump for older clients.

Overview
Grok Build is added as a third usage provider alongside Claude Code and Codex, so local Grok sessions under ~/.grok (or GROK_HOME) show up on the usage page.

The server scans ~/.grok/sessions/**/updates.jsonl only—skipping large chat_history / events logs—and parses turn_completed lines into per-model usage records with cache breakdowns and costUsdTicks converted to USD (10¹⁰ ticks per dollar). Multi-model turns get pro-rated cost when per-model ticks are missing or only partially reported. The usage contract adds grok to UsageProviderKind and bumps USAGE_CONTRACT_VERSION to 5; the scan cache accepts grok entries.

Web and mobile usage UI gain Grok Build labels, chart bands, icons, and theme-aware colors; the time breakdown table’s empty-state colSpan follows PROVIDER_ORDER so a third provider column doesn’t break layout.

^{Reviewed by Cursor Bugbot for commit 4c7a276. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Add Grok Build as a provider in the usage view

• Adds a parseGrokLine function in usageTranscripts.ts that parses Grok updates.jsonl session logs into per-model UsageRecord entries with cost tick-to-USD conversion and pro-rated cost attribution.
• Extends the server-side usage scanner in UsageService.ts to discover Grok sessions via $GROK_HOME or ~/.grok, filtering to updates.jsonl files only.
• Adds "grok" to UsageProviderKind and bumps the contract version from 4 to 5 in usage.ts.
• Updates mobile and web UI provider registries to display Grok Build with a label, color, and icon in charts and tables.
• Fixes the empty-state table colspan in UsagePage.tsx to compute dynamically from PROVIDER_ORDER.length instead of a hardcoded value.

^{Macroscope summarized 4c7a276.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 768538bd-fb54-4ea7-ae9f-9a73bed604b0

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

Approvability

Verdict: Needs human review

This PR introduces a new feature adding Grok Build as a third usage provider with substantial new parsing logic, contract version bump, and new user-facing behavior. New feature integrations warrant human review.

^{You can customize Macroscope's approvability policy. Learn more.}

[EliasofIX]

Submitting pending review so follow-up fix replies can land on the nits below.

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0d56431a655d962959d1899e171ec784c5b3c311. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

I am closing this because #8358 has landed in main and covers the same Grok historical usage reporting. I checked the current implementation and did not find any remaining change from this PR that we still need to merge.

Thank you for putting this together. The closure is only because the work is now covered elsewhere, not because there was a problem with your contribution.

[EliasofIX]

Summary

The usage page only scanned Claude Code and Codex session transcripts, so Grok Build work never showed up even though sessions under ~/.grok already carry per-turn token and cost data.

This adds Grok as a third usage provider end to end:

• Parse turn_completed usage from Grok updates.jsonl (per-model modelUsage, cache breakdown, costUsdTicks at 10^10 ticks per USD)
• Scan $GROK_HOME/sessions (default ~/.grok/sessions), restricted to updates.jsonl so large chat/events logs stay out of the cold path
• Surface Grok Build in the web and mobile usage charts/labels
• Bump the usage contract version for the new provider kind

Test plan

• Open Usage after pairing a local server that has Grok sessions in ~/.grok
• Confirm Grok Build appears in the provider split, daily chart, and model table
• Confirm Claude/Codex totals still match the previous behavior
• vp test run apps/server/src/usage/usageTranscripts.test.ts

Made with Grok Build (xAI).

---

Note

Medium Risk
New transcript parsing and cost pro-rating logic could mis-attribute tokens or dollars for edge-case Grok logs, but changes are confined to usage scanning/aggregation with broad test coverage and a contract version bump for older clients.

Overview
Grok Build is added as a third usage provider alongside Claude Code and Codex, so local Grok sessions under ~/.grok (or GROK_HOME) show up on the usage page.

The server scans ~/.grok/sessions/**/updates.jsonl only—skipping large chat_history / events logs—and parses turn_completed lines into per-model usage records with cache breakdowns and costUsdTicks converted to USD (10¹⁰ ticks per dollar). Multi-model turns get pro-rated cost when per-model ticks are missing or only partially reported. The usage contract adds grok to UsageProviderKind and bumps USAGE_CONTRACT_VERSION to 5; the scan cache accepts grok entries.

Web and mobile usage UI gain Grok Build labels, chart bands, icons, and theme-aware colors; the time breakdown table’s empty-state colSpan follows PROVIDER_ORDER so a third provider column doesn’t break layout.

^{Reviewed by Cursor Bugbot for commit 4c7a276. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Add Grok Build as a provider in the usage view

• Adds a parseGrokLine function in usageTranscripts.ts that parses Grok updates.jsonl session logs into per-model UsageRecord entries with cost tick-to-USD conversion and pro-rated cost attribution.
• Extends the server-side usage scanner in UsageService.ts to discover Grok sessions via $GROK_HOME or ~/.grok, filtering to updates.jsonl files only.
• Adds "grok" to UsageProviderKind and bumps the contract version from 4 to 5 in usage.ts.
• Updates mobile and web UI provider registries to display Grok Build with a label, color, and icon in charts and tables.
• Fixes the empty-state table colspan in UsagePage.tsx to compute dynamically from PROVIDER_ORDER.length instead of a hardcoded value.

^{Macroscope summarized 4c7a276.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 768538bd-fb54-4ea7-ae9f-9a73bed604b0

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

Approvability

Verdict: Needs human review

This PR introduces a new feature adding Grok Build as a third usage provider with substantial new parsing logic, contract version bump, and new user-facing behavior. New feature integrations warrant human review.

^{You can customize Macroscope's approvability policy. Learn more.}

[EliasofIX]

Submitting pending review so follow-up fix replies can land on the nits below.

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0d56431a655d962959d1899e171ec784c5b3c311. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

I am closing this because #8358 has landed in main and covers the same Grok historical usage reporting. I checked the current implementation and did not find any remaining change from this PR that we still need to merge.

Thank you for putting this together. The closure is only because the work is now covered elsewhere, not because there was a problem with your contribution.

[EliasofIX]

Summary

The usage page only scanned Claude Code and Codex session transcripts, so Grok Build work never showed up even though sessions under ~/.grok already carry per-turn token and cost data.

This adds Grok as a third usage provider end to end:

• Parse turn_completed usage from Grok updates.jsonl (per-model modelUsage, cache breakdown, costUsdTicks at 10^10 ticks per USD)
• Scan $GROK_HOME/sessions (default ~/.grok/sessions), restricted to updates.jsonl so large chat/events logs stay out of the cold path
• Surface Grok Build in the web and mobile usage charts/labels
• Bump the usage contract version for the new provider kind

Test plan

• Open Usage after pairing a local server that has Grok sessions in ~/.grok
• Confirm Grok Build appears in the provider split, daily chart, and model table
• Confirm Claude/Codex totals still match the previous behavior
• vp test run apps/server/src/usage/usageTranscripts.test.ts

Made with Grok Build (xAI).

---

Note

Medium Risk
New transcript parsing and cost pro-rating logic could mis-attribute tokens or dollars for edge-case Grok logs, but changes are confined to usage scanning/aggregation with broad test coverage and a contract version bump for older clients.

Overview
Grok Build is added as a third usage provider alongside Claude Code and Codex, so local Grok sessions under ~/.grok (or GROK_HOME) show up on the usage page.

The server scans ~/.grok/sessions/**/updates.jsonl only—skipping large chat_history / events logs—and parses turn_completed lines into per-model usage records with cache breakdowns and costUsdTicks converted to USD (10¹⁰ ticks per dollar). Multi-model turns get pro-rated cost when per-model ticks are missing or only partially reported. The usage contract adds grok to UsageProviderKind and bumps USAGE_CONTRACT_VERSION to 5; the scan cache accepts grok entries.

Web and mobile usage UI gain Grok Build labels, chart bands, icons, and theme-aware colors; the time breakdown table’s empty-state colSpan follows PROVIDER_ORDER so a third provider column doesn’t break layout.

^{Reviewed by Cursor Bugbot for commit 4c7a276. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Add Grok Build as a provider in the usage view

• Adds a parseGrokLine function in usageTranscripts.ts that parses Grok updates.jsonl session logs into per-model UsageRecord entries with cost tick-to-USD conversion and pro-rated cost attribution.
• Extends the server-side usage scanner in UsageService.ts to discover Grok sessions via $GROK_HOME or ~/.grok, filtering to updates.jsonl files only.
• Adds "grok" to UsageProviderKind and bumps the contract version from 4 to 5 in usage.ts.
• Updates mobile and web UI provider registries to display Grok Build with a label, color, and icon in charts and tables.
• Fixes the empty-state table colspan in UsagePage.tsx to compute dynamically from PROVIDER_ORDER.length instead of a hardcoded value.

^{Macroscope summarized 4c7a276.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 768538bd-fb54-4ea7-ae9f-9a73bed604b0

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

Approvability

Verdict: Needs human review

This PR introduces a new feature adding Grok Build as a third usage provider with substantial new parsing logic, contract version bump, and new user-facing behavior. New feature integrations warrant human review.

^{You can customize Macroscope's approvability policy. Learn more.}

[EliasofIX]

Submitting pending review so follow-up fix replies can land on the nits below.

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0d56431a655d962959d1899e171ec784c5b3c311. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

I am closing this because #8358 has landed in main and covers the same Grok historical usage reporting. I checked the current implementation and did not find any remaining change from this PR that we still need to merge.

Thank you for putting this together. The closure is only because the work is now covered elsewhere, not because there was a problem with your contribution.

[EliasofIX]

Summary

The usage page only scanned Claude Code and Codex session transcripts, so Grok Build work never showed up even though sessions under ~/.grok already carry per-turn token and cost data.

This adds Grok as a third usage provider end to end:

• Parse turn_completed usage from Grok updates.jsonl (per-model modelUsage, cache breakdown, costUsdTicks at 10^10 ticks per USD)
• Scan $GROK_HOME/sessions (default ~/.grok/sessions), restricted to updates.jsonl so large chat/events logs stay out of the cold path
• Surface Grok Build in the web and mobile usage charts/labels
• Bump the usage contract version for the new provider kind

Test plan

• Open Usage after pairing a local server that has Grok sessions in ~/.grok
• Confirm Grok Build appears in the provider split, daily chart, and model table
• Confirm Claude/Codex totals still match the previous behavior
• vp test run apps/server/src/usage/usageTranscripts.test.ts

Made with Grok Build (xAI).

---

Note

Medium Risk
New transcript parsing and cost pro-rating logic could mis-attribute tokens or dollars for edge-case Grok logs, but changes are confined to usage scanning/aggregation with broad test coverage and a contract version bump for older clients.

Overview
Grok Build is added as a third usage provider alongside Claude Code and Codex, so local Grok sessions under ~/.grok (or GROK_HOME) show up on the usage page.

The server scans ~/.grok/sessions/**/updates.jsonl only—skipping large chat_history / events logs—and parses turn_completed lines into per-model usage records with cache breakdowns and costUsdTicks converted to USD (10¹⁰ ticks per dollar). Multi-model turns get pro-rated cost when per-model ticks are missing or only partially reported. The usage contract adds grok to UsageProviderKind and bumps USAGE_CONTRACT_VERSION to 5; the scan cache accepts grok entries.

Web and mobile usage UI gain Grok Build labels, chart bands, icons, and theme-aware colors; the time breakdown table’s empty-state colSpan follows PROVIDER_ORDER so a third provider column doesn’t break layout.

^{Reviewed by Cursor Bugbot for commit 4c7a276. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Add Grok Build as a provider in the usage view

• Adds a parseGrokLine function in usageTranscripts.ts that parses Grok updates.jsonl session logs into per-model UsageRecord entries with cost tick-to-USD conversion and pro-rated cost attribution.
• Extends the server-side usage scanner in UsageService.ts to discover Grok sessions via $GROK_HOME or ~/.grok, filtering to updates.jsonl files only.
• Adds "grok" to UsageProviderKind and bumps the contract version from 4 to 5 in usage.ts.
• Updates mobile and web UI provider registries to display Grok Build with a label, color, and icon in charts and tables.
• Fixes the empty-state table colspan in UsagePage.tsx to compute dynamically from PROVIDER_ORDER.length instead of a hardcoded value.

^{Macroscope summarized 4c7a276.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 768538bd-fb54-4ea7-ae9f-9a73bed604b0

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

Approvability

Verdict: Needs human review

This PR introduces a new feature adding Grok Build as a third usage provider with substantial new parsing logic, contract version bump, and new user-facing behavior. New feature integrations warrant human review.

^{You can customize Macroscope's approvability policy. Learn more.}

[EliasofIX]

Submitting pending review so follow-up fix replies can land on the nits below.

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0d56431a655d962959d1899e171ec784c5b3c311. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

I am closing this because #8358 has landed in main and covers the same Grok historical usage reporting. I checked the current implementation and did not find any remaining change from this PR that we still need to merge.

Thank you for putting this together. The closure is only because the work is now covered elsewhere, not because there was a problem with your contribution.

[EliasofIX]

Summary

The usage page only scanned Claude Code and Codex session transcripts, so Grok Build work never showed up even though sessions under ~/.grok already carry per-turn token and cost data.

This adds Grok as a third usage provider end to end:

• Parse turn_completed usage from Grok updates.jsonl (per-model modelUsage, cache breakdown, costUsdTicks at 10^10 ticks per USD)
• Scan $GROK_HOME/sessions (default ~/.grok/sessions), restricted to updates.jsonl so large chat/events logs stay out of the cold path
• Surface Grok Build in the web and mobile usage charts/labels
• Bump the usage contract version for the new provider kind

Test plan

• Open Usage after pairing a local server that has Grok sessions in ~/.grok
• Confirm Grok Build appears in the provider split, daily chart, and model table
• Confirm Claude/Codex totals still match the previous behavior
• vp test run apps/server/src/usage/usageTranscripts.test.ts

Made with Grok Build (xAI).

---

Note

Medium Risk
New transcript parsing and cost pro-rating logic could mis-attribute tokens or dollars for edge-case Grok logs, but changes are confined to usage scanning/aggregation with broad test coverage and a contract version bump for older clients.

Overview
Grok Build is added as a third usage provider alongside Claude Code and Codex, so local Grok sessions under ~/.grok (or GROK_HOME) show up on the usage page.

The server scans ~/.grok/sessions/**/updates.jsonl only—skipping large chat_history / events logs—and parses turn_completed lines into per-model usage records with cache breakdowns and costUsdTicks converted to USD (10¹⁰ ticks per dollar). Multi-model turns get pro-rated cost when per-model ticks are missing or only partially reported. The usage contract adds grok to UsageProviderKind and bumps USAGE_CONTRACT_VERSION to 5; the scan cache accepts grok entries.

Web and mobile usage UI gain Grok Build labels, chart bands, icons, and theme-aware colors; the time breakdown table’s empty-state colSpan follows PROVIDER_ORDER so a third provider column doesn’t break layout.

^{Reviewed by Cursor Bugbot for commit 4c7a276. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Add Grok Build as a provider in the usage view

• Adds a parseGrokLine function in usageTranscripts.ts that parses Grok updates.jsonl session logs into per-model UsageRecord entries with cost tick-to-USD conversion and pro-rated cost attribution.
• Extends the server-side usage scanner in UsageService.ts to discover Grok sessions via $GROK_HOME or ~/.grok, filtering to updates.jsonl files only.
• Adds "grok" to UsageProviderKind and bumps the contract version from 4 to 5 in usage.ts.
• Updates mobile and web UI provider registries to display Grok Build with a label, color, and icon in charts and tables.
• Fixes the empty-state table colspan in UsagePage.tsx to compute dynamically from PROVIDER_ORDER.length instead of a hardcoded value.

^{Macroscope summarized 4c7a276.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 768538bd-fb54-4ea7-ae9f-9a73bed604b0

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

Approvability

Verdict: Needs human review

This PR introduces a new feature adding Grok Build as a third usage provider with substantial new parsing logic, contract version bump, and new user-facing behavior. New feature integrations warrant human review.

^{You can customize Macroscope's approvability policy. Learn more.}

[EliasofIX]

Submitting pending review so follow-up fix replies can land on the nits below.

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0d56431a655d962959d1899e171ec784c5b3c311. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

I am closing this because #8358 has landed in main and covers the same Grok historical usage reporting. I checked the current implementation and did not find any remaining change from this PR that we still need to merge.

Thank you for putting this together. The closure is only because the work is now covered elsewhere, not because there was a problem with your contribution.

[EliasofIX]

Summary

The usage page only scanned Claude Code and Codex session transcripts, so Grok Build work never showed up even though sessions under ~/.grok already carry per-turn token and cost data.

This adds Grok as a third usage provider end to end:

• Parse turn_completed usage from Grok updates.jsonl (per-model modelUsage, cache breakdown, costUsdTicks at 10^10 ticks per USD)
• Scan $GROK_HOME/sessions (default ~/.grok/sessions), restricted to updates.jsonl so large chat/events logs stay out of the cold path
• Surface Grok Build in the web and mobile usage charts/labels
• Bump the usage contract version for the new provider kind

Test plan

• Open Usage after pairing a local server that has Grok sessions in ~/.grok
• Confirm Grok Build appears in the provider split, daily chart, and model table
• Confirm Claude/Codex totals still match the previous behavior
• vp test run apps/server/src/usage/usageTranscripts.test.ts

Made with Grok Build (xAI).

---

Note

Medium Risk
New transcript parsing and cost pro-rating logic could mis-attribute tokens or dollars for edge-case Grok logs, but changes are confined to usage scanning/aggregation with broad test coverage and a contract version bump for older clients.

Overview
Grok Build is added as a third usage provider alongside Claude Code and Codex, so local Grok sessions under ~/.grok (or GROK_HOME) show up on the usage page.

The server scans ~/.grok/sessions/**/updates.jsonl only—skipping large chat_history / events logs—and parses turn_completed lines into per-model usage records with cache breakdowns and costUsdTicks converted to USD (10¹⁰ ticks per dollar). Multi-model turns get pro-rated cost when per-model ticks are missing or only partially reported. The usage contract adds grok to UsageProviderKind and bumps USAGE_CONTRACT_VERSION to 5; the scan cache accepts grok entries.

Web and mobile usage UI gain Grok Build labels, chart bands, icons, and theme-aware colors; the time breakdown table’s empty-state colSpan follows PROVIDER_ORDER so a third provider column doesn’t break layout.

^{Reviewed by Cursor Bugbot for commit 4c7a276. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Add Grok Build as a provider in the usage view

• Adds a parseGrokLine function in usageTranscripts.ts that parses Grok updates.jsonl session logs into per-model UsageRecord entries with cost tick-to-USD conversion and pro-rated cost attribution.
• Extends the server-side usage scanner in UsageService.ts to discover Grok sessions via $GROK_HOME or ~/.grok, filtering to updates.jsonl files only.
• Adds "grok" to UsageProviderKind and bumps the contract version from 4 to 5 in usage.ts.
• Updates mobile and web UI provider registries to display Grok Build with a label, color, and icon in charts and tables.
• Fixes the empty-state table colspan in UsagePage.tsx to compute dynamically from PROVIDER_ORDER.length instead of a hardcoded value.

^{Macroscope summarized 4c7a276.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 768538bd-fb54-4ea7-ae9f-9a73bed604b0

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

Approvability

Verdict: Needs human review

This PR introduces a new feature adding Grok Build as a third usage provider with substantial new parsing logic, contract version bump, and new user-facing behavior. New feature integrations warrant human review.

^{You can customize Macroscope's approvability policy. Learn more.}

[EliasofIX]

Submitting pending review so follow-up fix replies can land on the nits below.

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0d56431a655d962959d1899e171ec784c5b3c311. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

I am closing this because #8358 has landed in main and covers the same Grok historical usage reporting. I checked the current implementation and did not find any remaining change from this PR that we still need to merge.

Thank you for putting this together. The closure is only because the work is now covered elsewhere, not because there was a problem with your contribution.

[EliasofIX]

Summary

The usage page only scanned Claude Code and Codex session transcripts, so Grok Build work never showed up even though sessions under ~/.grok already carry per-turn token and cost data.

This adds Grok as a third usage provider end to end:

• Parse turn_completed usage from Grok updates.jsonl (per-model modelUsage, cache breakdown, costUsdTicks at 10^10 ticks per USD)
• Scan $GROK_HOME/sessions (default ~/.grok/sessions), restricted to updates.jsonl so large chat/events logs stay out of the cold path
• Surface Grok Build in the web and mobile usage charts/labels
• Bump the usage contract version for the new provider kind

Test plan

• Open Usage after pairing a local server that has Grok sessions in ~/.grok
• Confirm Grok Build appears in the provider split, daily chart, and model table
• Confirm Claude/Codex totals still match the previous behavior
• vp test run apps/server/src/usage/usageTranscripts.test.ts

Made with Grok Build (xAI).

---

Note

Medium Risk
New transcript parsing and cost pro-rating logic could mis-attribute tokens or dollars for edge-case Grok logs, but changes are confined to usage scanning/aggregation with broad test coverage and a contract version bump for older clients.

Overview
Grok Build is added as a third usage provider alongside Claude Code and Codex, so local Grok sessions under ~/.grok (or GROK_HOME) show up on the usage page.

The server scans ~/.grok/sessions/**/updates.jsonl only—skipping large chat_history / events logs—and parses turn_completed lines into per-model usage records with cache breakdowns and costUsdTicks converted to USD (10¹⁰ ticks per dollar). Multi-model turns get pro-rated cost when per-model ticks are missing or only partially reported. The usage contract adds grok to UsageProviderKind and bumps USAGE_CONTRACT_VERSION to 5; the scan cache accepts grok entries.

Web and mobile usage UI gain Grok Build labels, chart bands, icons, and theme-aware colors; the time breakdown table’s empty-state colSpan follows PROVIDER_ORDER so a third provider column doesn’t break layout.

^{Reviewed by Cursor Bugbot for commit 4c7a276. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Add Grok Build as a provider in the usage view

• Adds a parseGrokLine function in usageTranscripts.ts that parses Grok updates.jsonl session logs into per-model UsageRecord entries with cost tick-to-USD conversion and pro-rated cost attribution.
• Extends the server-side usage scanner in UsageService.ts to discover Grok sessions via $GROK_HOME or ~/.grok, filtering to updates.jsonl files only.
• Adds "grok" to UsageProviderKind and bumps the contract version from 4 to 5 in usage.ts.
• Updates mobile and web UI provider registries to display Grok Build with a label, color, and icon in charts and tables.
• Fixes the empty-state table colspan in UsagePage.tsx to compute dynamically from PROVIDER_ORDER.length instead of a hardcoded value.

^{Macroscope summarized 4c7a276.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 768538bd-fb54-4ea7-ae9f-9a73bed604b0

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

Approvability

Verdict: Needs human review

This PR introduces a new feature adding Grok Build as a third usage provider with substantial new parsing logic, contract version bump, and new user-facing behavior. New feature integrations warrant human review.

^{You can customize Macroscope's approvability policy. Learn more.}

[EliasofIX]

Submitting pending review so follow-up fix replies can land on the nits below.

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0d56431a655d962959d1899e171ec784c5b3c311. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

I am closing this because #8358 has landed in main and covers the same Grok historical usage reporting. I checked the current implementation and did not find any remaining change from this PR that we still need to merge.

Thank you for putting this together. The closure is only because the work is now covered elsewhere, not because there was a problem with your contribution.

[EliasofIX]

Summary

The usage page only scanned Claude Code and Codex session transcripts, so Grok Build work never showed up even though sessions under ~/.grok already carry per-turn token and cost data.

This adds Grok as a third usage provider end to end:

• Parse turn_completed usage from Grok updates.jsonl (per-model modelUsage, cache breakdown, costUsdTicks at 10^10 ticks per USD)
• Scan $GROK_HOME/sessions (default ~/.grok/sessions), restricted to updates.jsonl so large chat/events logs stay out of the cold path
• Surface Grok Build in the web and mobile usage charts/labels
• Bump the usage contract version for the new provider kind

Test plan

• Open Usage after pairing a local server that has Grok sessions in ~/.grok
• Confirm Grok Build appears in the provider split, daily chart, and model table
• Confirm Claude/Codex totals still match the previous behavior
• vp test run apps/server/src/usage/usageTranscripts.test.ts

Made with Grok Build (xAI).

---

Note

Medium Risk
New transcript parsing and cost pro-rating logic could mis-attribute tokens or dollars for edge-case Grok logs, but changes are confined to usage scanning/aggregation with broad test coverage and a contract version bump for older clients.

Overview
Grok Build is added as a third usage provider alongside Claude Code and Codex, so local Grok sessions under ~/.grok (or GROK_HOME) show up on the usage page.

The server scans ~/.grok/sessions/**/updates.jsonl only—skipping large chat_history / events logs—and parses turn_completed lines into per-model usage records with cache breakdowns and costUsdTicks converted to USD (10¹⁰ ticks per dollar). Multi-model turns get pro-rated cost when per-model ticks are missing or only partially reported. The usage contract adds grok to UsageProviderKind and bumps USAGE_CONTRACT_VERSION to 5; the scan cache accepts grok entries.

Web and mobile usage UI gain Grok Build labels, chart bands, icons, and theme-aware colors; the time breakdown table’s empty-state colSpan follows PROVIDER_ORDER so a third provider column doesn’t break layout.

^{Reviewed by Cursor Bugbot for commit 4c7a276. Bugbot is set up for automated code reviews on this repo. Configure here.}

Note

Add Grok Build as a provider in the usage view

• Adds a parseGrokLine function in usageTranscripts.ts that parses Grok updates.jsonl session logs into per-model UsageRecord entries with cost tick-to-USD conversion and pro-rated cost attribution.
• Extends the server-side usage scanner in UsageService.ts to discover Grok sessions via $GROK_HOME or ~/.grok, filtering to updates.jsonl files only.
• Adds "grok" to UsageProviderKind and bumps the contract version from 4 to 5 in usage.ts.
• Updates mobile and web UI provider registries to display Grok Build with a label, color, and icon in charts and tables.
• Fixes the empty-state table colspan in UsagePage.tsx to compute dynamically from PROVIDER_ORDER.length instead of a hardcoded value.

^{Macroscope summarized 4c7a276.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 768538bd-fb54-4ea7-ae9f-9a73bed604b0

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[macroscopeapp]

Approvability

Verdict: Needs human review

This PR introduces a new feature adding Grok Build as a third usage provider with substantial new parsing logic, contract version bump, and new user-facing behavior. New feature integrations warrant human review.

^{You can customize Macroscope's approvability policy. Learn more.}

[EliasofIX]

Submitting pending review so follow-up fix replies can land on the nits below.

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0d56431a655d962959d1899e171ec784c5b3c311. Configure here.}

[t3dotgg]

Note

🤖 GPT-5.6 Sol responding on behalf of Theo

I am closing this because #8358 has landed in main and covers the same Grok historical usage reporting. I checked the current implementation and did not find any remaining change from this PR that we still need to merge.

Thank you for putting this together. The closure is only because the work is now covered elsewhere, not because there was a problem with your contribution.