Release SDK updates

.github/workflows/stlc-promote.yml

@@ -69,6 +69,59 @@ jobs:
         env:
           GH_TOKEN: ${{ steps.prod-token.outputs.token }}
         run: |
+          # Regenerate the body on every promote. The release branch is
+          # force-pushed, so a body written only at gh-pr-create time freezes at
+          # whatever was pending the day the PR opened and never mentions
+          # anything promoted since.
+          git log --reverse --pretty=$'%H\t%s' production/main..origin/main > commits.tsv
+          PROD_REPO="${PRODUCTION_REPO}" python3 - commits.tsv > body.md <<'PY'
+          import os, re, sys
+
+          repo = os.environ["PROD_REPO"]
+          rows = [l.rstrip("\n").split("\t", 1) for l in open(sys.argv[1]) if "\t" in l]
+
+          NAMED = [("feat", "### Features"), ("fix", "### Bug Fixes"),
+                   ("perf", "### Performance"), ("docs", "### Documentation"),
+                   ("refactor", "### Refactors")]
+          CHORE = {"chore", "ci", "build", "test", "style", "release"}
+          NOISE = re.compile(r"^(Build SDK|Merge pull request #\d+|Merge branch )")
+
+          named, chores, unconventional, generated = {}, [], [], []
+          for sha, subject in rows:
+              m = re.match(r"^(\w+)(\([^)]*\))?!?:", subject)
+              kind = m.group(1) if m else ""
+              if kind in dict(NAMED):   named.setdefault(kind, []).append((sha, subject))
+              elif kind in CHORE:       chores.append((sha, subject))
+              elif NOISE.match(subject): generated.append((sha, subject))
+              else:                     unconventional.append((sha, subject))
+
+          def rows_md(items):
+              out = []
+              for sha, subject in items:
+                  # Five-digit refs are run-llama/platform PRs; GitHub would
+                  # otherwise autolink them to this repo's own issue numbers.
+                  subject = re.sub(r"\(#(\d{5,})\)",
+                                   r"([#\1](https://github.com/run-llama/platform/pull/\1))", subject)
+                  out.append(f"- {subject} ([`{sha[:7]}`](https://github.com/{repo}/commit/{sha}))")
+              return out
+
+          L = [f"Promotes `stainless/release` onto `main` — **{len(rows)} commits**. "
+               "Merging this lets release-please cut the next release and write `CHANGELOG.md`.", ""]
+          for kind, title in NAMED:
+              if named.get(kind):
+                  L += [title, *rows_md(named[kind]), ""]
+          if unconventional:
+              L += ["### :warning: Not conventional-commit — release-please will DROP these from `CHANGELOG.md`",
+                    *rows_md(unconventional), ""]
+          for items, title in ((chores, "Chores &amp; CI"), (generated, "Codegen &amp; merge commits")):
+              if items:
+                  L += [f"<details><summary>{title} ({len(items)})</summary>", "",
+                        *rows_md(items), "", "</details>", ""]
+          L.append("<sub>Regenerated by `stlc-promote.yml` on every promote, from "
+                   "`main...stainless/release`.</sub>")
+          print("\n".join(L))
+          PY
+
           EXISTING_PR=$(gh pr list \
             --repo "${PRODUCTION_REPO}" \
             --head stainless/release \
@@ -81,7 +134,9 @@ jobs:
               --base main \
               --head stainless/release \
               --title "Release SDK updates" \
-              --body "$(git log --oneline production/main..origin/main)"
+              --body-file body.md
           else
-            echo "Release PR #${EXISTING_PR} already exists. Force-push has updated it."
+            gh pr edit "${EXISTING_PR}" \
+              --repo "${PRODUCTION_REPO}" \
+              --body-file body.md
           fi

.github/workflows/stlc-sync-from-production.yml

@@ -1,15 +1,15 @@
 name: Sync from production
 
 on:
-  schedule:
-    # A poll, not a release — tune to how often releases/contributions land on
-    # production. Every few hours is plenty; daily is fine.
-    - cron: '17 */6 * * *'
+  # Cadence lives in run-llama/stainless .github/workflows/stlc-dispatch-backsync.yml,
+  # which dispatches this workflow with a CI_BOT App installation token. A `schedule:`
+  # here would fire it a second time and attribute that run to whoever last edited the
+  # cron. Verified 2026-08-07 by running both paths side by side: dispatched runs show
+  # llama-org-ci-bot[bot], scheduled runs showed a maintainer who had left the org.
   workflow_dispatch: {}
 
 permissions:
-  contents: write # push the back-sync branch on staging
-  pull-requests: write # open/merge the back-sync PR on staging
+  contents: read # checkout only — the push and the PR use the CI_BOT app token
 
 concurrency:
   group: stlc-sync-from-production
@@ -21,8 +21,6 @@ jobs:
     if: github.repository == 'run-llama/llamacloud-prod-java'
     env:
       PRODUCTION_REPO: run-llama/llama-parse-java
-      # The built-in token writes the branch + PR on this (staging) repo.
-      GH_TOKEN: ${{ github.token }}
     steps:
       - name: Generate production-repo token (CI_BOT GitHub App, read-only)
         id: prod-token
@@ -34,10 +32,24 @@ jobs:
           repositories: llama-parse-java
           permission-contents: read
 
+      - name: Generate staging-repo token (CI_BOT GitHub App)
+        id: staging-token
+        uses: actions/create-github-app-token@v1
+        with:
+          app-id: ${{ secrets.CI_BOT_APP_ID }}
+          private-key: ${{ secrets.CI_BOT_PRIVATE_KEY }}
+          # owner/repositories omitted: the token scopes to this repository.
+          permission-contents: write
+          permission-pull-requests: write
+          # production/main carries .github/workflows changes; GITHUB_TOKEN can
+          # never push those.
+          permission-workflows: write
+
       - name: Check out staging
         uses: actions/checkout@v6
         with:
           fetch-depth: 0
+          persist-credentials: false
 
       - name: Fetch production main
         env:
@@ -64,10 +76,17 @@ jobs:
 
       - name: Push production main to the back-sync branch on staging
         if: steps.diff.outputs.behind == 'true'
-        run: git push origin production/main:refs/heads/stlc/from-prod --force
+        env:
+          STAGING_REPO_TOKEN: ${{ steps.staging-token.outputs.token }}
+        run: |
+          git push \
+            "https://x-access-token:${STAGING_REPO_TOKEN}@github.com/${GITHUB_REPOSITORY}.git" \
+            production/main:refs/heads/stlc/from-prod --force
 
       - name: Open or update the back-sync PR on staging
         if: steps.diff.outputs.behind == 'true'
+        env:
+          GH_TOKEN: ${{ steps.staging-token.outputs.token }}
         run: |
           EXISTING_PR=$(gh pr list --head stlc/from-prod --base main --state open --json number --jq '.[0].number')
           if [ -z "${EXISTING_PR}" ]; then

.stats.yml

@@ -1 +1 @@
-configured_endpoints: 136
+configured_endpoints: 150

README.md

@@ -3,7 +3,7 @@
 <!-- x-release-please-start-version -->
 
 [![Maven Central](https://img.shields.io/maven-central/v/ai.llamaindex.llamacloud/llama-cloud)](https://central.sonatype.com/artifact/ai.llamaindex.llamacloud/llama-cloud/1.4.0)
-[![javadoc](https://javadoc.io/badge2/ai.llamaindex.llamacloud/llama-cloud/1.4.0/javadoc.svg)](https://javadoc.io/doc/ai.llamaindex.llamacloud/llama-cloud/1.3.0)
+[![javadoc](https://javadoc.io/badge2/ai.llamaindex.llamacloud/llama-cloud/1.4.0/javadoc.svg)](https://javadoc.io/doc/ai.llamaindex.llamacloud/llama-cloud/1.4.0)
 
 <!-- x-release-please-end -->
 

llama-cloud-core/src/main/kotlin/ai/llamaindex/llamacloud/client/LlamaCloudClient.kt

@@ -12,11 +12,15 @@ import ai.llamaindex.llamacloud.services.blocking.DataSinkService
 import ai.llamaindex.llamacloud.services.blocking.DataSourceService
 import ai.llamaindex.llamacloud.services.blocking.ExtractService
 import ai.llamaindex.llamacloud.services.blocking.FileService
+import ai.llamaindex.llamacloud.services.blocking.JobDataPointService
 import ai.llamaindex.llamacloud.services.blocking.ParsingService
 import ai.llamaindex.llamacloud.services.blocking.PipelineService
 import ai.llamaindex.llamacloud.services.blocking.ProjectService
 import ai.llamaindex.llamacloud.services.blocking.RetrieverService
 import ai.llamaindex.llamacloud.services.blocking.SheetService
+import ai.llamaindex.llamacloud.services.blocking.SplitService
+import ai.llamaindex.llamacloud.services.blocking.V2ProjectService
+import ai.llamaindex.llamacloud.services.blocking.WebhookConfigService
 import java.util.function.Consumer
 
 /**
@@ -59,6 +63,8 @@ interface LlamaCloudClient {
 
     fun sheets(): SheetService
 
+    fun split(): SplitService
+
     fun parsing(): ParsingService
 
     fun extract(): ExtractService
@@ -71,8 +77,14 @@ interface LlamaCloudClient {
 
     fun configurations(): ConfigurationService
 
+    fun webhookConfigs(): WebhookConfigService
+
     fun projects(): ProjectService
 
+    fun v2Projects(): V2ProjectService
+
+    fun jobDataPoints(): JobDataPointService
+
     fun dataSinks(): DataSinkService
 
     fun dataSources(): DataSourceService
@@ -110,6 +122,8 @@ interface LlamaCloudClient {
 
         fun sheets(): SheetService.WithRawResponse
 
+        fun split(): SplitService.WithRawResponse
+
         fun parsing(): ParsingService.WithRawResponse
 
         fun extract(): ExtractService.WithRawResponse
@@ -122,8 +136,14 @@ interface LlamaCloudClient {
 
         fun configurations(): ConfigurationService.WithRawResponse
 
+        fun webhookConfigs(): WebhookConfigService.WithRawResponse
+
         fun projects(): ProjectService.WithRawResponse
 
+        fun v2Projects(): V2ProjectService.WithRawResponse
+
+        fun jobDataPoints(): JobDataPointService.WithRawResponse
+
         fun dataSinks(): DataSinkService.WithRawResponse
 
         fun dataSources(): DataSourceService.WithRawResponse

llama-cloud-core/src/main/kotlin/ai/llamaindex/llamacloud/client/LlamaCloudClientAsync.kt

@@ -12,11 +12,15 @@ import ai.llamaindex.llamacloud.services.async.DataSinkServiceAsync
 import ai.llamaindex.llamacloud.services.async.DataSourceServiceAsync
 import ai.llamaindex.llamacloud.services.async.ExtractServiceAsync
 import ai.llamaindex.llamacloud.services.async.FileServiceAsync
+import ai.llamaindex.llamacloud.services.async.JobDataPointServiceAsync
 import ai.llamaindex.llamacloud.services.async.ParsingServiceAsync
 import ai.llamaindex.llamacloud.services.async.PipelineServiceAsync
 import ai.llamaindex.llamacloud.services.async.ProjectServiceAsync
 import ai.llamaindex.llamacloud.services.async.RetrieverServiceAsync
 import ai.llamaindex.llamacloud.services.async.SheetServiceAsync
+import ai.llamaindex.llamacloud.services.async.SplitServiceAsync
+import ai.llamaindex.llamacloud.services.async.V2ProjectServiceAsync
+import ai.llamaindex.llamacloud.services.async.WebhookConfigServiceAsync
 import java.util.function.Consumer
 
 /**
@@ -59,6 +63,8 @@ interface LlamaCloudClientAsync {
 
     fun sheets(): SheetServiceAsync
 
+    fun split(): SplitServiceAsync
+
     fun parsing(): ParsingServiceAsync
 
     fun extract(): ExtractServiceAsync
@@ -71,8 +77,14 @@ interface LlamaCloudClientAsync {
 
     fun configurations(): ConfigurationServiceAsync
 
+    fun webhookConfigs(): WebhookConfigServiceAsync
+
     fun projects(): ProjectServiceAsync
 
+    fun v2Projects(): V2ProjectServiceAsync
+
+    fun jobDataPoints(): JobDataPointServiceAsync
+
     fun dataSinks(): DataSinkServiceAsync
 
     fun dataSources(): DataSourceServiceAsync
@@ -114,6 +126,8 @@ interface LlamaCloudClientAsync {
 
         fun sheets(): SheetServiceAsync.WithRawResponse
 
+        fun split(): SplitServiceAsync.WithRawResponse
+
         fun parsing(): ParsingServiceAsync.WithRawResponse
 
         fun extract(): ExtractServiceAsync.WithRawResponse
@@ -126,8 +140,14 @@ interface LlamaCloudClientAsync {
 
         fun configurations(): ConfigurationServiceAsync.WithRawResponse
 
+        fun webhookConfigs(): WebhookConfigServiceAsync.WithRawResponse
+
         fun projects(): ProjectServiceAsync.WithRawResponse
 
+        fun v2Projects(): V2ProjectServiceAsync.WithRawResponse
+
+        fun jobDataPoints(): JobDataPointServiceAsync.WithRawResponse
+
         fun dataSinks(): DataSinkServiceAsync.WithRawResponse
 
         fun dataSources(): DataSourceServiceAsync.WithRawResponse