docs: add Bedrock runbooks for Claude Code and Codex - #5868
Conversation
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (1)
🚧 Files skipped from review as they are similar to previous changes (1)
📝 WalkthroughSummary by CodeRabbit
WalkthroughThis PR adds Claude Code and Codex CLI runbooks for connecting to Amazon Bedrock through Bifrost. It also adds an Integrations “Runbooks” navigation group that links to both pages. ChangesBedrock CLI runbooks
Estimated code review effort: 2 (Simple) | ~10 minutes 🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
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. Comment |
There was a problem hiding this comment.
Actionable comments posted: 6
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@docs/runbooks/claude-code-bedrock.mdx`:
- Around line 21-26: Update the Claude Haiku entry in the model mapping table to
use the configured deployment name, changing the Direct Claude Code model value
to bedrock/claude-haiku-4-5-20251001 so it matches the models or aliases
configuration referenced elsewhere in the runbook.
- Around line 45-77: Update the Bedrock configuration example to use the
schema-supported bedrock_key_config.deployments mapping instead of aliases,
matching the model-to-deployment structure defined in
transports/config.schema.json. If aliases must remain, explicitly state the
minimum Bifrost version and provide the corresponding version-specific
configuration.
- Around line 95-100: Update the virtual-key configuration instructions in the
“Configure the virtual key” section to require selecting the Bedrock provider
key in key_ids and adding the intended bare deployment names to allowed_models.
State that empty allowlists deny access, and document ["*"] only as the explicit
allow-all value while preserving the existing active, budget, and rate-limit
checks.
In `@docs/runbooks/codex-bedrock.mdx`:
- Around line 33-35: Update the provider references in the Warning block so the
supported provider remains labeled “AWS Bedrock” and the provider users must
avoid is labeled with its actual classic-provider name. Preserve the surrounding
model ID and deployment-name guidance.
- Around line 102-105: Update the virtual-key setup guidance in the
codex-bedrock runbook so “Allow the bedrock provider” includes the actual
provider config allowlists required for access: specify the intended deployment
names in allowed_models and the provider key IDs in key_ids, or explicitly use
"*" only when unrestricted access is intended. Keep the existing
active/budget/rate-limit step, and anchor the edit around the virtual key
allowlist instructions so the documented path matches the provider config
requirement.
- Around line 50-89: Update the JSON example under the “Aliases” section to use
the schema-defined Bedrock deployment mapping at bedrock_key_config.deployments
instead of generic aliases, mapping each model name to its Bedrock model ID
according to transports/config.schema.json. Preserve any static alias
configuration separately only if the documented setup requires it, and ensure
the example matches the schema’s supported structure.
🪄 Autofix
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro Plus
Run ID: 296547f5-b5bd-4f53-bad4-9b4805b3e2ed
⛔ Files ignored due to path filters (1)
docs/media/ui-bedrock-deployment-mappings.pngis excluded by!**/*.png
📒 Files selected for processing (3)
docs/docs.jsondocs/runbooks/claude-code-bedrock.mdxdocs/runbooks/codex-bedrock.mdx
There was a problem hiding this comment.
Actionable comments posted: 3
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@docs/runbooks/claude-code-bedrock.mdx`:
- Around line 57-64: Update the bedrock_key_config example to include the
required Bedrock region property defined by transports/config.schema.json,
placing it alongside the deployments configuration and using the schema’s exact
property name and structure so the JSON is valid copy-paste configuration.
In `@docs/runbooks/codex-bedrock.mdx`:
- Line 37: Update the deployment-name guidance in the Codex Bedrock runbook to
distinguish the unqualified Bifrost deployment name, such as gpt-5.6-sol, from
the provider-qualified Codex model identifier, such as bedrock/gpt-5.6-sol.
Clarify that Codex should select the provider-qualified value while the
deployment configuration retains the unqualified name, preserving the required
exact matching relationship.
- Around line 63-71: Update the bedrock_key_config example to include the
required Bedrock region setting alongside deployments, using the configuration
key and region value expected by the Bedrock provider documentation.
🪄 Autofix
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro Plus
Run ID: bcc683ce-f9b4-4970-9e37-dfb355bd878f
📒 Files selected for processing (2)
docs/runbooks/claude-code-bedrock.mdxdocs/runbooks/codex-bedrock.mdx
There was a problem hiding this comment.
Actionable comments posted: 1
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (5)
docs/runbooks/claude-code-bedrock.mdx (4)
84-88: 🔒 Security & Privacy | 🟠 Major | ⚡ Quick winKeep
ANTHROPIC_AUTH_TOKENout of project settings.
docs/runbooks/claude-code-bedrock.mdxsays to edit.claude/settings.json, then showsANTHROPIC_AUTH_TOKENin both example snippets. Project settings can be shared or checked into source control, while.claude/settings.local.jsonis personal. Store the virtual key in~/.claude/settings.jsonor.claude/settings.local.jsononly, and removeANTHROPIC_AUTH_TOKENfrom the project-scoped instructions.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/runbooks/claude-code-bedrock.mdx` around lines 84 - 88, Update the Claude Code Bedrock instructions and example snippets to remove ANTHROPIC_AUTH_TOKEN from project-scoped settings, keeping it only in user-level ~/.claude/settings.json or personal .claude/settings.local.json examples. Preserve other project settings and configuration guidance.Source: Path instructions
90-90: 🎯 Functional Correctness | 🟡 Minor | ⚡ Quick winCorrect the model precedence statement.
ANTHROPIC_MODELtakes precedence over the Claude Codemodelsetting, so the current wording is opposite to Claude Code precedence.Proposed wording
-Remove a top-level `model` setting if one is present because it overrides the environment-based model selection below. +If a top-level `model` setting is present, remove it only when it conflicts with the intended model; `ANTHROPIC_MODEL` takes precedence over that setting.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/runbooks/claude-code-bedrock.mdx` at line 90, Correct the model precedence statement in the runbook: state that ANTHROPIC_MODEL takes precedence over Claude Code’s top-level model setting, and remove the instruction claiming the model setting overrides environment-based selection.
127-127: 🎯 Functional Correctness | 🟡 Minor | ⚡ Quick winMention Claude Code cache refresh before using
/model.Gateway model discovery is cached locally and refreshed at startup. If users add or change the discovery setting in an existing Claude Code session, tell them to restart Claude Code before running
/model.Proposed wording
-Once that's done, you can call `/model` primitive on your claude code. That will show you list of models configured on Bifrost for the virtual key. +Restart Claude Code after changing the file. Then run the `/model` command to view the models returned by Bifrost for the virtual key.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/runbooks/claude-code-bedrock.mdx` at line 127, Update the Claude Code gateway discovery instructions around the `/model` guidance to state that discovery results are cached and refreshed at startup. Tell users to restart an existing Claude Code session after adding or changing the discovery setting, then run `/model` to view the configured models.
129-131: 🎯 Functional Correctness | 🟡 Minor | ⚡ Quick winVersion the discovered model ID filter.
The
bedrock/claude-sonnet-...andbedrock/anthropic.claude-sonnet-...IDs pass discovery in Claude Code 2.1.223 and later because the filter acceptsclaudeoranthropicanywhere in the ID, case-insensitively. Versions before 2.1.223 require the ID to begin withclaudeoranthropic. Split the warning by version: newer versions still ignore IDs without those substrings anywhere, such ascoding-sonnet.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/runbooks/claude-code-bedrock.mdx` around lines 129 - 131, Update the Warning in the Claude Code Bedrock runbook to distinguish behavior by Claude Code version: document that versions before 2.1.223 require discovered IDs to begin with “claude” or “anthropic,” while 2.1.223 and later accept those terms anywhere case-insensitively but still ignore IDs such as “coding-sonnet.”docs/runbooks/codex-bedrock.mdx (1)
149-149: 🎯 Functional Correctness | 🟡 Minor | ⚡ Quick winUse a cross-platform catalog path example.
The paragraph uses
~/.codex/bifrost_catalog.json, but the TOML example points to/Users/<you>/..., which only works on macOS. Update the example to match the catalog location for the README target OS, or provide platform-specific paths for Linux/Windows.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@docs/runbooks/codex-bedrock.mdx` at line 149, Update the catalog path examples in the deployment instructions so they are consistent across the prose and TOML configuration and work on the README’s target platform; otherwise provide clearly labeled Linux, macOS, and Windows path variants. Keep the referenced bifrost_catalog.json location and surrounding setup guidance unchanged.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@docs/runbooks/claude-code-bedrock.mdx`:
- Around line 146-152: Update the validation instructions in the Claude Code
runbook to branch on discovery status: when disabled, direct users to /model,
select a built-in tier, and send the validation prompt; when enabled, allow
selecting either a displayed **From gateway** deployment or the matching
built-in tier before sending the prompt.
---
Outside diff comments:
In `@docs/runbooks/claude-code-bedrock.mdx`:
- Around line 84-88: Update the Claude Code Bedrock instructions and example
snippets to remove ANTHROPIC_AUTH_TOKEN from project-scoped settings, keeping it
only in user-level ~/.claude/settings.json or personal
.claude/settings.local.json examples. Preserve other project settings and
configuration guidance.
- Line 90: Correct the model precedence statement in the runbook: state that
ANTHROPIC_MODEL takes precedence over Claude Code’s top-level model setting, and
remove the instruction claiming the model setting overrides environment-based
selection.
- Line 127: Update the Claude Code gateway discovery instructions around the
`/model` guidance to state that discovery results are cached and refreshed at
startup. Tell users to restart an existing Claude Code session after adding or
changing the discovery setting, then run `/model` to view the configured models.
- Around line 129-131: Update the Warning in the Claude Code Bedrock runbook to
distinguish behavior by Claude Code version: document that versions before
2.1.223 require discovered IDs to begin with “claude” or “anthropic,” while
2.1.223 and later accept those terms anywhere case-insensitively but still
ignore IDs such as “coding-sonnet.”
In `@docs/runbooks/codex-bedrock.mdx`:
- Line 149: Update the catalog path examples in the deployment instructions so
they are consistent across the prose and TOML configuration and work on the
README’s target platform; otherwise provide clearly labeled Linux, macOS, and
Windows path variants. Keep the referenced bifrost_catalog.json location and
surrounding setup guidance unchanged.
🪄 Autofix
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro Plus
Run ID: eb56e441-8cea-40b5-b109-36d938e42813
📒 Files selected for processing (2)
docs/runbooks/claude-code-bedrock.mdxdocs/runbooks/codex-bedrock.mdx
|
Note GitHub couldn't provide a complete incremental comparison for this pull request, so CodeRabbit is performing a full review instead. This review may take a little longer. |
* docs: add Bedrock runbooks for Claude Code and Codex * docs: use Bedrock deployment mappings in runbooks * remove unecessary warning * replace static json with UI image * docs: add Edge setup paths to Bedrock runbooks * docs: clarify Claude Code model validation
* docs: add Bedrock runbooks for Claude Code and Codex * docs: use Bedrock deployment mappings in runbooks * remove unecessary warning * replace static json with UI image * docs: add Edge setup paths to Bedrock runbooks * docs: clarify Claude Code model validation
* docs: add Bedrock runbooks for Claude Code and Codex * docs: use Bedrock deployment mappings in runbooks * remove unecessary warning * replace static json with UI image * docs: add Edge setup paths to Bedrock runbooks * docs: clarify Claude Code model validation
* docs: add Bedrock runbooks for Claude Code and Codex * docs: use Bedrock deployment mappings in runbooks * remove unecessary warning * replace static json with UI image * docs: add Edge setup paths to Bedrock runbooks * docs: clarify Claude Code model validation
* docs: add Bedrock runbooks for Claude Code and Codex * docs: use Bedrock deployment mappings in runbooks * remove unecessary warning * replace static json with UI image * docs: add Edge setup paths to Bedrock runbooks * docs: clarify Claude Code model validation
* docs: add Bedrock runbooks for Claude Code and Codex * docs: use Bedrock deployment mappings in runbooks * remove unecessary warning * replace static json with UI image * docs: add Edge setup paths to Bedrock runbooks * docs: clarify Claude Code model validation
Summary
Validation
make docsand reviewed the Mintlify preview locally