tensorzero · virajmehta · Dec 26, 2025 · Dec 25, 2025 · Dec 25, 2025 · Dec 25, 2025
diff --git a/.cargo/config.toml b/.cargo/config.toml
@@ -105,8 +105,6 @@ test-feedback-load = [
 
 build-e2e = "build --bin gateway --features e2e_tests"
 run-e2e = "run --bin gateway --features e2e_tests -- --config-file tensorzero-core/tests/e2e/config/tensorzero.*.toml"
-run-e2e-mock-batch = "run --bin gateway --features e2e_tests -- --config-file tensorzero-core/tests/e2e/config/{tensorzero.*.toml,mock_batch.toml}"
-run-e2e-mock-optimization = "run --bin gateway --features e2e_tests -- --config-file tensorzero-core/tests/e2e/config/{tensorzero.*.toml,mock_optimization.toml}"
 migrate-postgres = "run --bin gateway --features e2e_tests -- --run-postgres-migrations"
 watch-e2e = "watch -x run-e2e"
 

diff --git a/.github/workflows/build-gateway-e2e-container.yml b/.github/workflows/build-gateway-e2e-container.yml
@@ -53,9 +53,9 @@ jobs:
 
       # For some reason, 'docker compose build --push' doesn't work when using a remote builder (i.e. Namespace)
       - name: Build test containers
-        run: docker compose -f tensorzero-core/tests/e2e/docker-compose.live.yml build --push mock-inference-provider provider-proxy gateway live-tests
+        run: docker compose -f tensorzero-core/tests/e2e/docker-compose.live.yml build --push provider-proxy gateway live-tests
 
       # Note that this pushes an e2e build of the gateway to 'tensorzero/gateway-e2e'.
       # It does *not* push to the production 'tensorzero/gateway' repo.
       - name: Push test containers
-        run: docker compose -f tensorzero-core/tests/e2e/docker-compose.live.yml push mock-inference-provider provider-proxy gateway live-tests
+        run: docker compose -f tensorzero-core/tests/e2e/docker-compose.live.yml push provider-proxy gateway live-tests
diff --git a/...kflows/build-mock-inference-container.yml → ...ows/build-mock-provider-api-container.yml b/...kflows/build-mock-inference-container.yml → ...ows/build-mock-provider-api-container.yml
@@ -1,10 +1,10 @@
-name: Build Mock Inference Container
+name: Build Mock Provider API Container
 
 on:
   workflow_call:
 
 jobs:
-  build-mock-inference-container:
+  build-mock-provider-api-container:
     runs-on: ubuntu-latest
     if: github.repository == 'tensorzero/tensorzero'
     permissions:
@@ -30,18 +30,18 @@ jobs:
           wait-for-builder: true
         continue-on-error: ${{ github.event.pull_request.head.repo.full_name != github.repository || github.actor == 'dependabot[bot]' }}
 
-      - name: Build `mock-inference-provider` container
+      - name: Build `mock-provider-api` container
         run: |
-          docker buildx build --build-arg BUILDKIT_CONTEXT_KEEP_GIT_DIR=1 -f tensorzero-core/tests/mock-inference-provider/Dockerfile . -t tensorzero/mock-inference-provider:sha-${{ github.sha }} -t nscr.io/igvf4asmf8kri/mock-inference-provider:sha-${{ github.sha }}
+          docker buildx build --build-arg BUILDKIT_CONTEXT_KEEP_GIT_DIR=1 -f tensorzero-core/tests/mock-provider-api/Dockerfile . -t tensorzero/mock-provider-api:sha-${{ github.sha }} -t nscr.io/igvf4asmf8kri/mock-provider-api:sha-${{ github.sha }}
 
-      - name: Save `mock-inference-provider` container
-        run: docker save tensorzero/mock-inference-provider:sha-${{ github.sha }} > mock-inference-container.tar
+      - name: Save `mock-provider-api` container
+        run: docker save tensorzero/mock-provider-api:sha-${{ github.sha }} > mock-provider-api-container.tar
 
-      - name: Upload `mock-inference-provider` container as an artifact
+      - name: Upload `mock-provider-api` container as an artifact
         uses: actions/upload-artifact@330a01c490aca151604b8cf639adc76d48f6c5d4
         with:
-          name: build-mock-inference-container
-          path: mock-inference-container.tar
+          name: build-mock-provider-api-container
+          path: mock-provider-api-container.tar
           retention-days: 1
           if-no-files-found: error
           overwrite: false
@@ -50,6 +50,6 @@ jobs:
         run: nsc docker login
         continue-on-error: ${{ github.event.pull_request.head.repo.full_name != github.repository || github.actor == 'dependabot[bot]' }}
 
-      - name: Push `mock-inference` container to Namespace registry
-        run: docker push nscr.io/igvf4asmf8kri/mock-inference-provider:sha-${{ github.sha }}
+      - name: Push `mock-provider-api` container to Namespace registry
+        run: docker push nscr.io/igvf4asmf8kri/mock-provider-api:sha-${{ github.sha }}
         continue-on-error: ${{ github.event.pull_request.head.repo.full_name != github.repository || github.actor == 'dependabot[bot]' }}
diff --git a/.github/workflows/general.yml b/.github/workflows/general.yml
@@ -734,7 +734,7 @@ jobs:
       contents: read
       # Permission to download artifacts and for rust-cache
       actions: read
-    needs: [build-gateway-container, build-mock-inference-container]
+    needs: [build-gateway-container, build-mock-provider-api-container]
 
     # We don't run many tests here, so use a normal runner with Github Actions caching
     # to avoid unnecessarily using Namespace credits (it should still always finish before
@@ -803,10 +803,10 @@ jobs:
       - name: Download container images
         run: |
           docker pull nscr.io/igvf4asmf8kri/gateway:sha-${{ github.sha }}
-          docker pull nscr.io/igvf4asmf8kri/mock-inference-provider:sha-${{ github.sha }}
+          docker pull nscr.io/igvf4asmf8kri/mock-provider-api:sha-${{ github.sha }}
           # Retag the images to what we expect the names to be
           docker tag nscr.io/igvf4asmf8kri/gateway:sha-${{ github.sha }} tensorzero/gateway:sha-${{ github.sha }}
-          docker tag nscr.io/igvf4asmf8kri/mock-inference-provider:sha-${{ github.sha }} tensorzero/mock-inference-provider:sha-${{ github.sha }}
+          docker tag nscr.io/igvf4asmf8kri/mock-provider-api:sha-${{ github.sha }} tensorzero/mock-provider-api:sha-${{ github.sha }}
         continue-on-error: ${{ github.event.pull_request.head.repo.full_name != github.repository || github.actor == 'dependabot[bot]' }}
 
       - name: Download ClickHouse fixtures
@@ -834,9 +834,9 @@ jobs:
         run: |
           echo "TENSORZERO_GATEWAY_TAG=sha-${{ github.sha }}" >> $GITHUB_ENV
 
-      - name: Set TENSORZERO_MOCK_INFERENCE_PROVIDER_TAG
+      - name: Set TENSORZERO_MOCK_PROVIDER_API_TAG
         run: |
-          echo "TENSORZERO_MOCK_INFERENCE_PROVIDER_TAG=sha-${{ github.sha }}" >> $GITHUB_ENV
+          echo "TENSORZERO_MOCK_PROVIDER_API_TAG=sha-${{ github.sha }}" >> $GITHUB_ENV
 
       - name: Launch dependency services with non-replicated ClickHouse container for E2E tests
         if: matrix.replicated == false
@@ -945,7 +945,7 @@ jobs:
         if: always()
         run: cat e2e_logs.txt
 
-  # Run 'cargo test-optimization' against mock-inference-provider
+  # Run 'cargo test-optimization' against mock-provider-api
   mock-optimization-tests:
     permissions:
       contents: read
@@ -956,9 +956,8 @@ jobs:
     env:
       OPENAI_API_KEY: not_used
       FIREWORKS_API_KEY: not_used
-      FIREWORKS_ACCOUNT_ID: not_used
       TOGETHER_API_KEY: not_used
-      TENSORZERO_USE_MOCK_INFERENCE_PROVIDER: 1
+      TENSORZERO_INTERNAL_MOCK_PROVIDER_API: http://localhost:3030
       TENSORZERO_SKIP_LARGE_FIXTURES: 1
       R2_ACCESS_KEY_ID: ${{ secrets.R2_ACCESS_KEY_ID }}
       R2_SECRET_ACCESS_KEY: ${{ secrets.R2_SECRET_ACCESS_KEY }}
@@ -1012,7 +1011,7 @@ jobs:
 
       - name: Launch the gateway for E2E tests
         run: |
-          cargo run-e2e-mock-optimization > e2e_logs.txt 2>&1 &
+          cargo run-e2e > e2e_logs.txt 2>&1 &
           GATEWAY_PID=$!
           while ! curl -s http://localhost:3000/health; do
             if ! kill -0 $GATEWAY_PID 2>/dev/null; then
@@ -1058,8 +1057,8 @@ jobs:
       # Permission to fetch GitHub OIDC token authentication
       id-token: write
 
-  build-mock-inference-container:
-    uses: ./.github/workflows/build-mock-inference-container.yml
+  build-mock-provider-api-container:
+    uses: ./.github/workflows/build-mock-provider-api-container.yml
     permissions:
       # Permission to checkout the repository
       contents: read
@@ -1073,7 +1072,7 @@ jobs:
     uses: ./.github/workflows/ui-tests.yml
     with:
       is_merge_group: ${{ github.event_name == 'merge_group' }}
-    needs: [build-gateway-container, build-mock-inference-container]
+    needs: [build-gateway-container, build-mock-provider-api-container]
 
   ui-tests-e2e:
     permissions:
@@ -1086,7 +1085,7 @@ jobs:
       [
         build-gateway-container,
         build-ui-container,
-        build-mock-inference-container,
+        build-mock-provider-api-container,
       ]
     secrets:
       AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
@@ -1150,7 +1149,7 @@ jobs:
       [
         build-gateway-container,
         build-ui-container,
-        build-mock-inference-container,
+        build-mock-provider-api-container,
       ]
     secrets:
       S3_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
@@ -1167,7 +1166,7 @@ jobs:
       [
         build-gateway-container,
         build-gateway-e2e-container,
-        build-mock-inference-container,
+        build-mock-provider-api-container,
       ]
     if: (github.repository == 'tensorzero/tensorzero' && github.event_name == 'merge_group')
     permissions:
@@ -1191,7 +1190,7 @@ jobs:
       actions: read
     if: github.repository == 'tensorzero/tensorzero' && github.event_name == 'merge_group'
     uses: ./.github/workflows/mocked-batch-test.yml
-    needs: [build-mock-inference-container]
+    needs: [build-mock-provider-api-container]
     secrets: inherit
 
   # See 'ci/README.md' at the repository root for more details.
@@ -1212,7 +1211,7 @@ jobs:
         build-ui-container,
         build-gateway-container,
         build-gateway-e2e-container,
-        build-mock-inference-container,
+        build-mock-provider-api-container,
         mocked-batch-tests,
         minikube,
         rust-build,

diff --git a/.github/workflows/merge-queue.yml b/.github/workflows/merge-queue.yml
@@ -156,6 +156,8 @@ jobs:
       contents: read
       # Permission to fetch GitHub OIDC token authentication
       id-token: write
+      # Permission to download artifacts
+      actions: read
     timeout-minutes: 45
     strategy:
       matrix:
@@ -315,11 +317,11 @@ jobs:
       - name: Install Python for python async client tests
         run: uv python install 3.9
 
-      - name: "Python: PyO3 Client: pytest"
+      - name: "Python: PyO3 Client: pytest (non-mock tests)"
         working-directory: clients/python
         run: |
           # Start the test in background and capture its PID
-          bash ./test.sh --verbose -n 8 &
+          bash ./test.sh --verbose -n 8 -m "not mock" &
           TEST_PID=$!
           echo "Started test.sh with PID: $TEST_PID"
 
@@ -358,6 +360,13 @@ jobs:
           fi
           exit 1
 
+      - name: "Python: PyO3 Client: pytest (mock tests)"
+        working-directory: clients/python
+        env:
+          TENSORZERO_INTERNAL_MOCK_PROVIDER_API: http://localhost:3030
+        run: |
+          bash ./test.sh --verbose -n 8 -m mock
+
       - name: "Node.js: OpenAI Client: test"
         working-directory: clients/openai-node
         run: |

diff --git a/.github/workflows/mocked-batch-test.yml b/.github/workflows/mocked-batch-test.yml
@@ -11,7 +11,6 @@ env:
   AZURE_OPENAI_DEPLOYMENT_ID: "fake_deployment_id"
   DEEPSEEK_API_KEY: "fake_deepseek_key"
   FIREWORKS_API_KEY: "fake_fireworks_key"
-  FIREWORKS_ACCOUNT_ID: "fake_fireworks_account"
   FORCE_COLOR: 1
   GCP_VERTEX_CREDENTIALS_PATH: ${{ github.workspace }}/gcp_jwt_key.json
   GOOGLE_APPLICATION_CREDENTIALS: ${{ github.workspace }}/gcp_jwt_key.json
@@ -28,7 +27,7 @@ env:
   VLLM_API_BASE: "http://fake-vllm-endpoint:8000"
   VLLM_MODEL_NAME: "microsoft/Phi-3.5-mini-instruct"
   XAI_API_KEY: "fake_xai_key"
-  TENSORZERO_USE_MOCK_INFERENCE_PROVIDER: 1
+  TENSORZERO_INTERNAL_MOCK_PROVIDER_API: http://localhost:3030
 
 on:
   workflow_dispatch:
@@ -81,35 +80,35 @@ jobs:
           GCP_JWT_KEY: ${{ secrets.GCP_JWT_KEY }}
         run: echo "$GCP_JWT_KEY" > $GITHUB_WORKSPACE/gcp_jwt_key.json
 
-      - name: Download mock-inference-container
+      - name: Download mock-provider-api-container
         continue-on-error: true # Skip if missing (e.g. triggered via `workflow_dispatch`)
         uses: actions/download-artifact@018cc2cf5baa6db3ef3c5f8a56943fffe632ef53
         with:
-          name: build-mock-inference-container
+          name: build-mock-provider-api-container
           path: .
 
-      - name: Load mock-inference-container
-        id: load-mock-inference
+      - name: Load mock-provider-api-container
+        id: load-mock-provider-api
         continue-on-error: true # Skip if missing (e.g. triggered via `workflow_dispatch`)
         run: |
-          docker load < mock-inference-container.tar
+          docker load < mock-provider-api-container.tar
 
-      - name: Set TENSORZERO_MOCK_INFERENCE_PROVIDER_TAG
-        if: steps.load-mock-inference.outcome == 'success' # Only run if artifact was successfully loaded
+      - name: Set TENSORZERO_MOCK_PROVIDER_API_TAG
+        if: steps.load-mock-provider-api.outcome == 'success' # Only run if artifact was successfully loaded
         run: |
-          echo "TENSORZERO_MOCK_INFERENCE_PROVIDER_TAG=sha-${{ github.sha }}" >> $GITHUB_ENV
+          echo "TENSORZERO_MOCK_PROVIDER_API_TAG=sha-${{ github.sha }}" >> $GITHUB_ENV
 
-      - name: Start Docker services (mock-inference-provider + dependencies)
+      - name: Start Docker services (mock-provider-api + dependencies)
         run: |
-          docker compose -f tensorzero-core/tests/e2e/docker-compose.yml up -d --wait clickhouse postgres gateway-clickhouse-migrations gateway-postgres-migrations minio mock-inference-provider
+          docker compose -f tensorzero-core/tests/e2e/docker-compose.yml up -d --wait clickhouse postgres gateway-clickhouse-migrations gateway-postgres-migrations minio mock-provider-api
 
       - name: Set up TENSORZERO_CLICKHOUSE_URL for batch tests
         run: |
           echo "TENSORZERO_CLICKHOUSE_URL=http://chuser:chpassword@localhost:8123/tensorzero_e2e_tests" >> $GITHUB_ENV
 
       - name: Launch the gateway for batch tests
         run: |
-          cargo run-e2e-mock-batch > batch_logs.txt 2>&1 &
+          cargo run-e2e > batch_logs.txt 2>&1 &
           echo "GATEWAY_PID=$!" >> $GITHUB_ENV
           while ! curl -s -f http://localhost:3000/health >/dev/null 2>&1; do
             echo "Waiting for gateway to be healthy..."

diff --git a/.github/workflows/slash-command-regen-fixtures.yml b/.github/workflows/slash-command-regen-fixtures.yml
@@ -54,19 +54,17 @@ jobs:
           # Set real OpenAI and S3 keys here, so that we can run image evaluations to regenerate fixtures.
           # If we ever need to use any other providers to regenerate the fixtures, set those keys here.
           echo "ANTHROPIC_API_KEY=${{ secrets.ANTHROPIC_API_KEY }}" >> ui/fixtures/.env
-          echo "FIREWORKS_ACCOUNT_ID=fake_fireworks_account" >> ui/fixtures/.env
           echo "FIREWORKS_API_KEY=not_used" >> ui/fixtures/.env
-          echo "FIREWORKS_BASE_URL=http://mock-inference-provider:3030/fireworks/" >> ui/fixtures/.env
           echo "OPENAI_API_KEY=${{ secrets.OPENAI_API_KEY }}" >> ui/fixtures/.env
-          echo "OPENAI_BASE_URL=http://mock-inference-provider:3030/openai/" >> ui/fixtures/.env
           echo "S3_ACCESS_KEY_ID=${{ secrets.AWS_ACCESS_KEY_ID }}" >> ui/fixtures/.env
           echo "S3_SECRET_ACCESS_KEY=${{ secrets.AWS_SECRET_ACCESS_KEY }}" >> ui/fixtures/.env
           echo "TENSORZERO_CLICKHOUSE_URL=http://chuser:chpassword@clickhouse:8123/tensorzero_ui_fixtures" >> ui/fixtures/.env
           echo "TENSORZERO_GATEWAY_TAG=sha-${{ github.sha }}" >> ui/fixtures/.env
           echo "TENSORZERO_GATEWAY_URL=http://gateway:3000" >> ui/fixtures/.env
+          echo "TENSORZERO_INTERNAL_MOCK_PROVIDER_API=http://mock-provider-api:3030" >> ui/fixtures/.env
+          echo "TENSORZERO_MOCK_PROVIDER_API_TAG=sha-${{ github.sha }}" >> ui/fixtures/.env
           echo "TENSORZERO_UI_TAG=sha-${{ github.sha }}" >> ui/fixtures/.env
           echo "TOGETHER_API_KEY=not_used" >> ui/fixtures/.env
-          echo "TOGETHER_BASE_URL=http://mock-inference-provider:3030/together/" >> ui/fixtures/.env
           ./ui/fixtures/regenerate-model-inference-cache.sh
 
       - name: Upload to R2 and update download script

diff --git a/.github/workflows/ui-tests-e2e-model-inference-cache.yml b/.github/workflows/ui-tests-e2e-model-inference-cache.yml
@@ -88,7 +88,7 @@ jobs:
       - name: Download container images
         uses: actions/download-artifact@018cc2cf5baa6db3ef3c5f8a56943fffe632ef53
         with:
-          pattern: build-{gateway,ui,mock-inference}-container
+          pattern: build-{gateway,ui,mock-provider-api}-container
           merge-multiple: true
 
       - name: Load container images
@@ -98,7 +98,7 @@ jobs:
         run: |
           docker load < gateway-container.tar
           docker load < ui-container.tar
-          docker load < mock-inference-container.tar
+          docker load < mock-provider-api-container.tar
 
       # This allows us to use 'no-build' on subsequent steps
       - name: Build needed docker images
@@ -111,16 +111,11 @@ jobs:
         run: |
           # Environment variables shared by the gateway and ui containers
           echo "TENSORZERO_CLICKHOUSE_URL=http://chuser:chpassword@clickhouse:8123/tensorzero_ui_fixtures" >> fixtures/.env
-          echo "TENSORZERO_GATEWAY_URL=http://gateway:3000" >> fixtures/.env
           echo "TENSORZERO_GATEWAY_TAG=sha-${{ github.sha }}" >> fixtures/.env
+          echo "TENSORZERO_GATEWAY_URL=http://gateway:3000" >> fixtures/.env
+          echo "TENSORZERO_INTERNAL_MOCK_PROVIDER_API=http://mock-provider-api:3030" >> fixtures/.env
+          echo "TENSORZERO_MOCK_PROVIDER_API_TAG=sha-${{ github.sha }}" >> fixtures/.env
           echo "TENSORZERO_UI_TAG=sha-${{ github.sha }}" >> fixtures/.env
-          echo "TENSORZERO_MOCK_INFERENCE_PROVIDER_TAG=sha-${{ github.sha }}" >> fixtures/.env
-          # We need these set in the ui container, so that we construct the correct optimizer config
-          # to pass to 'experimentalLaunchOptimizationWorkflow'
-          echo "FIREWORKS_BASE_URL=http://mock-inference-provider:3030/fireworks/" >> fixtures/.env
-          echo "OPENAI_BASE_URL=http://mock-inference-provider:3030/openai/" >> fixtures/.env
-          echo "TOGETHER_BASE_URL=http://mock-inference-provider:3030/together/" >> fixtures/.env
-          echo "FIREWORKS_ACCOUNT_ID=fake_fireworks_account" >> fixtures/.env
           echo "VITE_TENSORZERO_FORCE_CACHE_ON=1" >> fixtures/.env
 
       - name: Enable authentication in config