base · montycheese · Aug 14, 2026
diff --git a/.cursor/rules/project-structure.mdc b/.cursor/rules/project-structure.mdc
@@ -11,23 +11,9 @@ alwaysApply: false
 - **benchmark/**: Go-based benchmark runner and core logic
 - **contracts/**: Solidity smart contracts for testing
 - **configs/**: YAML configuration files for different benchmark scenarios
-- **report/**: React-based web UI for viewing benchmark results
 - **runner/**: Go-based execution engine
 - **scripts/**: Utility scripts for setup and automation
 
-### Report UI Structure ([report/](mdc:report))
-```
-report/
-├── src/
-│   ├── components/     # Reusable UI components
-│   ├── pages/         # Page-level components
-│   ├── hooks/         # Custom React hooks
-│   ├── utils/         # Utility functions and formatters
-│   └── types.ts       # TypeScript type definitions
-├── public/            # Static assets
-└── package.json       # Frontend dependencies
-```
-
 ## Configuration Management
 
 ### Benchmark Configs ([configs/](mdc:configs))
@@ -44,12 +30,6 @@ report/
 
 ## Development Workflow
 
-### Frontend Development
-1. Work in the `report/` directory for UI changes
-2. Use TypeScript for type safety
-3. Follow React functional component patterns
-4. Use Tailwind CSS for styling
-
 ### Backend Development
 1. Work in `benchmark/` and `runner/` directories
 2. Use Go for performance-critical code
@@ -64,12 +44,6 @@ report/
 
 ## Build and Deployment
 
-### Frontend Build
-- Vite-based build system
-- Static asset generation
-- Optimized bundle output
-- Development server with hot reload
-
 ### Backend Build
 - Go module-based dependencies
 - Cross-platform compilation
@@ -82,5 +56,6 @@ report/
 1. Configuration files define test parameters
 2. Go runner executes benchmarks
 3. Results stored in structured format
-4. React UI displays and analyzes results
+   (see [docs/report-data-contract.md](mdc:docs/report-data-contract.md))
+4. The benchmark UI in `base/ui` reads that data and displays it
 
diff --git a/.github/assets/run-switcher.png b/.github/assets/run-switcher.png
diff --git a/.github/workflows/examples.yaml b/.github/workflows/examples.yaml
@@ -86,24 +86,9 @@ jobs:
 
           ./.github/scripts/run-example-benchmarks.sh ${{ runner.temp }}
 
-      - name: Build Report
-        run: |
-          cp -r ${{ runner.temp }}/output/ ./output/
-          pushd report
-          npm install
-          npm run build
-          popd
-
       - name: Upload Output
         uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02
         if: always()
         with:
           name: output
           path: ${{ runner.temp }}/output/
-
-      # upload to github actions artifact
-      - name: Upload Report
-        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02
-        with:
-          name: report
-          path: report/dist/
diff --git a/.github/workflows/nodejs.yaml b/.github/workflows/nodejs.yaml
diff --git a/.github/workflows/public-benchmarks.yaml b/.github/workflows/public-benchmarks.yaml
@@ -81,36 +81,12 @@ jobs:
             --reth-bin ${{ runner.temp }}/bin/reth \
             --geth-bin ${{ runner.temp }}/bin/geth
 
-      - name: Build Report
-        run: |
-          cp -r ${{ runner.temp }}/output/ ./output/
-          pushd report
-          npm install
-          npm run build
-          popd
-
-      - name: Upload static files as artifact
-        id: deployment
-        uses: actions/upload-pages-artifact@56afc609e74202658d3ffba0e8f6dda462b719fa # v3.0.1
+      # The report UI that used to be built here and published to GitHub Pages
+      # was removed from this repo; the run output is uploaded as a workflow
+      # artifact instead. See docs/report-data-contract.md for the schema.
+      - name: Upload Output
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02
+        if: always()
         with:
-          path: report/dist/
-
-  deploy:
-    runs-on: ubuntu-latest
-    needs: basic-benchmarks
-    permissions:
-      pages: write
-      id-token: write
-
-    environment:
-      name: github-pages
-      url: ${{ steps.deployment.outputs.page_url }}
-
-    steps:
-      - name: Harden the runner (Audit all outbound calls)
-        uses: step-security/harden-runner@002fdce3c6a235733a90a27c80493a3241e56863 # v2.12.1
-        with:
-          egress-policy: audit
-
-      - name: Publish Report
-        uses: actions/deploy-pages@d6db90164ac5ed86f2b6aed7e0febac5b3c0c03e # v4.0.5
+          name: output
+          path: ${{ runner.temp }}/output/
diff --git a/Makefile b/Makefile
@@ -53,11 +53,3 @@ build-base-reth-node:
 
 .PHONY: build-binaries
 build-binaries: build-reth build-geth build-base-reth-node
-
-.PHONY: build-frontend
-build-frontend:
-	cd report && yarn build
-
-.PHONY: run-frontend
-run-frontend:
-	cd report && set -a && [ -f ../.env ] && . ../.env && set +a && yarn dev
diff --git a/README.md b/README.md
@@ -27,21 +27,20 @@ Base Benchmark is a performance testing framework for Ethereum execution clients
 
 ## Results
 
-Public results are available at the following links:
+Benchmark runs write JSON to an output directory (see
+[docs/report-data-contract.md](docs/report-data-contract.md) for the schema).
+The `run-publish-benchmarks` workflow runs the public benchmark suite on every
+push to `main` and uploads its output as a workflow artifact.
 
-| Network      | Link                                                                   |
-| ------------ | ---------------------------------------------------------------------- |
-| Devnet       | [https://base.github.io/benchmark/](https://base.github.io/benchmark/) |
-| Base Sepolia | Coming soon                                                            |
-| Base Mainnet | Coming soon                                                            |
+This repo no longer ships a dashboard for viewing that output; it is a
+benchmark runner and a data contract.
 
 ## Features
 
 - **Performance Evaluation:** Test both block building and validation performance across execution clients (Geth, Reth, and more)
 - **Comparative Analysis:** Measure client behavior across various inputs and workloads
 - **Metric Collection:** Track critical metrics including submission times, latency, and throughput
 - **Flexible Workloads:** Configure transaction patterns to match your specific needs
-- **Interactive Dashboard:** Generate beautiful HTML reports with charts and run comparisons
 - **Import & Merge:** Combine benchmark results from multiple machines with flexible tagging
 
 ## Repository Structure
@@ -65,16 +64,13 @@ Public results are available at the following links:
 │   └── public/           # Production-ready benchmarks
 ├── contracts/            # Smart contracts for testing
 │   └── src/              # Solidity source files
-├── report/               # Interactive dashboard
-│   └── src/              # React TypeScript application
 └── clients/              # Client build scripts
 ```
 
 ## Prerequisites
 
 - **Go:** Version 1.21 or later. Install from [go.dev](https://go.dev/dl/)
 - **Foundry:** For smart contract compilation. See [installation guide](https://book.getfoundry.sh/getting-started/installation)
-- **Node.js:** Version 18+ for the interactive dashboard. Install from [nodejs.org](https://nodejs.org/)
 
 ## Getting Started
 
@@ -119,15 +115,12 @@ To see available options:
 ./bin/base-bench run --help
 ```
 
-### 5. View Results in the Interactive Dashboard
+### 5. Read the Results
 
-```bash
-cd report/
-npm install
-npm run dev
-```
-
-Open your browser to the URL shown (typically `http://localhost:5173`).
+The run writes one directory per benchmark under `--output-dir`, each holding a
+`metadata.json` and per-role `metrics-<role>.json` timeseries.
+[docs/report-data-contract.md](docs/report-data-contract.md) documents the
+layout and every field, so you can consume it from your own tooling.
 
 ## Available Benchmarks
 
@@ -241,11 +234,9 @@ OPTIONS:
 
 ### Understanding Runs and Suites
 
-When you view benchmark results in the interactive dashboard, you can switch between different test runs using the run switcher:
-
-<div align="center">
-  <img src=".github/assets/run-switcher.png" alt="Run Switcher" width="600">
-</div>
+A suite is one execution of `base-bench run`. Each suite writes its own set of
+runs to the output directory, keyed by the `BenchmarkRun` ID they share, so
+results from separate executions stay distinguishable.
 
 ### Creating Test Runs
 

diff --git a/configs/README.md b/configs/README.md
@@ -101,4 +101,6 @@ All configurations track key metrics:
 - **Resource Usage**: CPU, memory, disk I/O
 - **Throughput**: Transactions per second, gas per second
 
-For detailed metric analysis, view the interactive dashboard after running benchmarks.
+Every metric is written to the output directory as JSON. See
+[../docs/report-data-contract.md](../docs/report-data-contract.md) for the
+layout and field definitions.
diff --git a/docs/report-data-contract.md b/docs/report-data-contract.md
@@ -1,10 +1,10 @@
 # Report data contract
 
-This document describes what the `report/` web UI in this repo expects
-from whatever produces benchmark data. The intent is to let `base/benchmark`
-become a **pure visualization tool** — any runner (Go, Rust, future
-monorepo) can produce data the report consumes, as long as it follows
-this contract.
+This document describes what the benchmark report UI expects from whatever
+produces benchmark data. The UI itself lives in [`base/ui`](https://github.com/base/ui)
+under `app/benchmark/`; this repo is one of the producers. Any runner (Go,
+Rust, future monorepo) can produce data the report consumes, as long as it
+follows this contract.
 
 The current Go runner in `runner/` is one producer; the future Rust
 runner in the Coinbase monorepo will be another. Both write the same
@@ -191,7 +191,7 @@ on-the-wire shape.
 
 The set of metric keys inside `ExecutionMetrics` is open. The
 frontend has a registry of "known" metrics with units and labels at
-`report/src/metricDefinitions.ts`; anything not listed there still
+`base/ui/app/benchmark/metricDefinitions.ts`; anything not listed there still
 renders, just with raw key names.
 
 ## Comparison groups (automatic)
@@ -294,7 +294,7 @@ A natural follow-on is to drop the merged `metadata.json` response
 entirely and serve per-run JSON via an index endpoint
 (`GET /api/v1/runs?since=<ts>` paginated). The frontend already has a
 data-source abstraction in
-`report/src/services/dataService.ts` — swapping out the metadata
+`base/ui/app/benchmark/services/dataService.ts` — swapping out the metadata
 fetch would be a single function change. But this requires a
 frontend change, so it's a bigger lift than the producer-side
 upgrades above.
@@ -336,11 +336,11 @@ UI silently breaks.
 - Comparison synthesizer:
   `protocols/base-benchmarking/report-api/internal/services/comparison.go`
 - Frontend types (what the UI expects):
-  `report/src/types.ts::BenchmarkRun`
+  `base/ui/app/benchmark/types.ts::BenchmarkRun`
 - Frontend metric definitions (keys/units/labels):
-  `report/src/metricDefinitions.ts`
+  `base/ui/app/benchmark/metricDefinitions.ts`
 - Frontend filter logic (how dropdowns are auto-discovered):
-  `report/src/hooks/useBenchmarkFilters.ts` and `report/src/filter.ts`
+  `base/ui/app/benchmark/hooks/useBenchmarkFilters.ts` and `base/ui/app/benchmark/filter.ts`
 - Current Go runner that produces this shape:
   `runner/benchmark/result_metadata.go::Run` and
   `runner/service.go::applyClientVersion`