- Add max_tokens field to ModelProviderInfo struct
- Read from provider config, default to 8192 if not set
- Add debug logging for max_tokens value
- Allows per-provider customization in config.toml
Example config:
```toml
[model_providers.litellm]
max_tokens = 16384
```
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Regenerated lockfile with compatible dependency versions
- Resolved tree-sitter-language version conflict
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Add cache_control to system instructions (stable, high value)
- Add cache_control to conversation history (4 messages before end)
- Implements full Anthropic caching hierarchy: tools → system → history
- Significant cost savings for repeated similar requests
Cache hierarchy:
1. Tools (last tool) - already implemented
2. System instructions - new
3. Conversation history - new (stable portion)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Changed tracing from 0.1.51 to 0.1.41
- Regenerated Cargo.lock with available dependencies
- Version 0.1.51 does not exist on crates.io
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Fix orphaned tool_use error by tracking skip state per call_id
- Handle retries with same call_id correctly
- Add comprehensive debug logging for troubleshooting
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Fix: Skip empty/whitespace text content blocks
- Fix: Validate function call arguments and skip malformed calls
- Fix: Skip outputs for skipped function calls to maintain consistency
- Resolves Anthropic API errors:
- "messages: text content blocks must contain non-whitespace text"
- "Extra data: line 1 column 26 (char 25)" (invalid JSON)
- "unexpected `tool_use_id` found in `tool_result` blocks"
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Fix: Skip empty/whitespace-only text content blocks in chat completions
- This resolves Anthropic API errors:
- "messages: text content blocks must contain non-whitespace text"
- "messages: text content blocks must be non-empty"
- Updated version strings in all test files and snapshots
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Added scope: "@valknarthing" to setup-node action (required for scoped packages)
- Added npm update step to ensure npm CLI v11.5.1+ (required for OIDC support)
Matches the original OpenAI Codex workflow configuration.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Removed NODE_AUTH_TOKEN env var from publish-npm job
OIDC/Trusted Publishers authentication doesn't need NPM_TOKEN secret
- Commented out update-branch job since latest-alpha-cli branch doesn't exist
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Updated all version references from 0.1.1 to 0.1.2:
- Workspace version in llmx-rs/Cargo.toml
- Package version in llmx-cli/package.json
- Updated Cargo.lock with all workspace crate versions
- Updated test hardcoded version strings in:
- mcp-server/tests/common/mcp_process.rs
- app-server/tests/suite/user_agent.rs
- app-server/tests/common/mcp_process.rs
- Updated TUI snapshot tests with new version number
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Fixed environment variable names from LITELLM_* to LLMX_*
- LITELLM_BASE_URL → LLMX_BASE_URL
- LITELLM_API_KEY → LLMX_API_KEY
- Updated all documentation links to use absolute GitHub URLs instead of relative paths
- Fixes 404 errors when viewing README on npm registry
- All ./docs/ and ./LITELLM-SETUP.md links now point to github.com/valknarthing/llmx
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Add id-token: write permission for OIDC authentication
- Add --provenance flag to npm publish for supply chain security
- Use NODE_AUTH_TOKEN environment variable (set by setup-node)
- Remove manual .npmrc token writing (handled by setup-node with OIDC)
This enables automated npm publishing without storing tokens as secrets.
Requires Trusted Publisher to be configured at:
https://www.npmjs.com/package/@valknarthing/llmx/access🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Scoped packages (@valknarthing/llmx) require the --access public flag
for all npm publish commands, not just the first one.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Update hardcoded version strings in test files to match the new 0.1.1 version:
- llmx-app-server user_agent test
- llmx-app-server test helper initialize
- llmx-mcp-server test helper initialize
- llmx-tui status snapshot tests (auto-updated via cargo insta)
All previously failing tests now pass.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Update project branding to reflect community fork status:
- Changed welcome message from "OpenAI's command-line coding agent"
to "your command-line coding agent"
- Updated system prompt from "led by OpenAI" to "community project"
- Removed broken developers.openai.com/llmx documentation URLs
- Updated Windows setup instructions with correct npm package name
- Removed OpenAI CLA (not applicable to community fork)
- Removed OpenAI open source fund documentation
- Updated config docs to remove managed configuration references
- Changed npm package reference from @openai/llmx to @valknarthing/llmx
These changes complete the rebranding from OpenAI's project to an
independent community fork while maintaining proper attribution to
the original project in README.md.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Use read_stream_until_response_message instead of private read_jsonrpc_message
- Simplify test to accept both success and error outcomes for cancel
- Remove unused imports (JSONRPCMessage, CancelLoginChatGptResponse)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Remove .github/*.png images (not needed)
- Remove Homebrew installation instructions (no cask available)
- Fix original project URL to point to openai/codex
- Fix GitHub URLs from valknar to valknarthing
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Delete existing release before creating new one to avoid conflicts
- Remove llmx-responses-api-proxy from dotslash config (not needed)
- Add 'Create GitHub Release' step to upload platform binaries
- Configure npm authentication by writing token to setup-node's .npmrc
- Use internal package identifier 'llmx' in workflow scripts
This fixes the release workflow to properly:
1. Clean up previous failed releases
2. Create GitHub release with all platform binaries
3. Add DotSlash manifest file (llmx) without conflicts
4. Publish to npm as @valknarthing/llmx with proper authentication
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
The test was failing when the login session was cleaned up before the cancel
request could be processed. Now the test accepts both successful cancellation
and 'login id not found' error as valid outcomes.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Revert workflow to use internal identifier 'llmx' instead of scoped name
- Revert build_npm_package.py to only accept internal identifiers
- Restore "Create GitHub Release" step that was incorrectly removed
- Package name '@valknarthing/llmx' is read from package.json by the script
This matches OpenAI's approach where they use 'codex' not '@openai/codex' in workflow.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
The build_npm_package.py script now accepts '@valknarthing/llmx' as a valid
--package argument and normalizes it to 'llmx' internally for processing.
This allows the workflow to use the full scoped package name for clarity
while maintaining backward compatibility with the 'llmx' identifier.
The 'Create GitHub Release' step was conflicting with dotslash-publish-release.
Both were trying to upload assets to the same release, causing the 'asset under
the same name already exists' error.
Removed the softprops/action-gh-release step since facebook/dotslash-publish-release
handles creating the release and uploading all assets including the dotslash manifest.
The dotslash-publish-release step was failing because the 'Create GitHub Release'
step had already uploaded the binary assets. Adding clobber: true allows dotslash
to overwrite/reuse existing assets when creating the dotslash manifest.
Changed from npm Trusted Publishing (OIDC) to traditional token authentication.
The NODE_AUTH_TOKEN environment variable now uses the NPM_TOKEN secret.
Removed:
- OIDC id-token permission
- npm CLI update step (no longer needed)
Added:
- NODE_AUTH_TOKEN in both Setup Node.js and Publish to npm steps
The stage_npm_packages.py script creates tarballs named {package}-npm-{version}.tgz
(e.g., llmx-npm-0.1.0.tgz), but the workflow was looking for valknar-llmx-npm-{version}.tgz.
Fixed the download pattern and publish tarball list to match the actual output from
the staging script.
The build_npm_package.py script only accepts specific package choices:
'llmx', 'llmx-responses-api-proxy', 'llmx-sdk'. The scoped package name
'@valknar/llmx' is not a valid choice.
The package name 'llmx' is the correct identifier used internally by the
build system, which then produces the scoped npm package.
The script was failing when creating temporary directories for scoped packages like @valknar/llmx
because the forward slash in the package name was being used directly in the temp directory prefix,
causing Python's tempfile.mkdtemp() to fail with 'No such file or directory'.
Fix by sanitizing the package name: replace '/' with '-' and remove '@' before using it in the
temp directory prefix. The actual package output file still uses the original package name.
Fixes: FileNotFoundError in GitHub Actions rust-release workflow
This release represents a comprehensive transformation of the codebase from Codex to LLMX,
enhanced with LiteLLM integration to support 100+ LLM providers through a unified API.
## Major Changes
### Phase 1: Repository & Infrastructure Setup
- Established new repository structure and branching strategy
- Created comprehensive project documentation (CLAUDE.md, LITELLM-SETUP.md)
- Set up development environment and tooling configuration
### Phase 2: Rust Workspace Transformation
- Renamed all Rust crates from `codex-*` to `llmx-*` (30+ crates)
- Updated package names, binary names, and workspace members
- Renamed core modules: codex.rs → llmx.rs, codex_delegate.rs → llmx_delegate.rs
- Updated all internal references, imports, and type names
- Renamed directories: codex-rs/ → llmx-rs/, codex-backend-openapi-models/ → llmx-backend-openapi-models/
- Fixed all Rust compilation errors after mass rename
### Phase 3: LiteLLM Integration
- Integrated LiteLLM for multi-provider LLM support (Anthropic, OpenAI, Azure, Google AI, AWS Bedrock, etc.)
- Implemented OpenAI-compatible Chat Completions API support
- Added model family detection and provider-specific handling
- Updated authentication to support LiteLLM API keys
- Renamed environment variables: OPENAI_BASE_URL → LLMX_BASE_URL
- Added LLMX_API_KEY for unified authentication
- Enhanced error handling for Chat Completions API responses
- Implemented fallback mechanisms between Responses API and Chat Completions API
### Phase 4: TypeScript/Node.js Components
- Renamed npm package: @codex/codex-cli → @valknar/llmx
- Updated TypeScript SDK to use new LLMX APIs and endpoints
- Fixed all TypeScript compilation and linting errors
- Updated SDK tests to support both API backends
- Enhanced mock server to handle multiple API formats
- Updated build scripts for cross-platform packaging
### Phase 5: Configuration & Documentation
- Updated all configuration files to use LLMX naming
- Rewrote README and documentation for LLMX branding
- Updated config paths: ~/.codex/ → ~/.llmx/
- Added comprehensive LiteLLM setup guide
- Updated all user-facing strings and help text
- Created release plan and migration documentation
### Phase 6: Testing & Validation
- Fixed all Rust tests for new naming scheme
- Updated snapshot tests in TUI (36 frame files)
- Fixed authentication storage tests
- Updated Chat Completions payload and SSE tests
- Fixed SDK tests for new API endpoints
- Ensured compatibility with Claude Sonnet 4.5 model
- Fixed test environment variables (LLMX_API_KEY, LLMX_BASE_URL)
### Phase 7: Build & Release Pipeline
- Updated GitHub Actions workflows for LLMX binary names
- Fixed rust-release.yml to reference llmx-rs/ instead of codex-rs/
- Updated CI/CD pipelines for new package names
- Made Apple code signing optional in release workflow
- Enhanced npm packaging resilience for partial platform builds
- Added Windows sandbox support to workspace
- Updated dotslash configuration for new binary names
### Phase 8: Final Polish
- Renamed all assets (.github images, labels, templates)
- Updated VSCode and DevContainer configurations
- Fixed all clippy warnings and formatting issues
- Applied cargo fmt and prettier formatting across codebase
- Updated issue templates and pull request templates
- Fixed all remaining UI text references
## Technical Details
**Breaking Changes:**
- Binary name changed from `codex` to `llmx`
- Config directory changed from `~/.codex/` to `~/.llmx/`
- Environment variables renamed (CODEX_* → LLMX_*)
- npm package renamed to `@valknar/llmx`
**New Features:**
- Support for 100+ LLM providers via LiteLLM
- Unified authentication with LLMX_API_KEY
- Enhanced model provider detection and handling
- Improved error handling and fallback mechanisms
**Files Changed:**
- 578 files modified across Rust, TypeScript, and documentation
- 30+ Rust crates renamed and updated
- Complete rebrand of UI, CLI, and documentation
- All tests updated and passing
**Dependencies:**
- Updated Cargo.lock with new package names
- Updated npm dependencies in llmx-cli
- Enhanced OpenAPI models for LLMX backend
This release establishes LLMX as a standalone project with comprehensive LiteLLM
integration, maintaining full backward compatibility with existing functionality
while opening support for a wide ecosystem of LLM providers.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Co-Authored-By: Sebastian Krüger <support@pivoine.art>
Unified exec isn't working on Linux because we don't provide the correct
arg0.
The library we use for pty management doesn't allow setting arg0
separately from executable. Use the same aliasing strategy we use for
`apply_patch` for `codex-linux-sandbox`.
Use `#[ctor]` hack to dispatch codex-linux-sandbox calls.
Addresses https://github.com/openai/codex/issues/6450
This PR is to unlock future WinGet installation. WinGet struggles to
create command aliases when installing from nested ZIPs on some clients,
so adding raw Windows x64/Arm64 executables lets the manifest use
InstallerType: portable with direct EXEs, which reliably registers the
codex alias. This makes “winget install → codex” work out of the box
without PATH changes.
---------
Co-authored-by: Michael Bolin <mbolin@openai.com>
We already do this for notification definitions and it's really nice.
Verified there are no changes to actual exported files by diff'ing
before and after this change.
- Moved the unix-only suspend/resume logic into a dedicated job_control
module housing SuspendContext, replacing scattered cfg-gated fields and
helpers in tui.rs.
- Tui now holds a single suspend_context (Arc-backed) instead of
multiple atomics, and the event stream uses it directly for Ctrl-Z
handling.
- Added detailed docs around the suspend/resume flow, cursor tracking,
and the Arc/atomic ownership model for the 'static event stream.
- Renamed the process-level SIGTSTP helper to suspend_process and the
cursor tracker to set_cursor_y to better reflect their roles.
This adds a debugging tool for analyzing why certain commands fail to
execute under the sandbox.
Example output:
```
$ codex debug seatbelt --log-denials bash -lc "(echo foo > ~/foo.txt)"
bash: /Users/nornagon/foo.txt: Operation not permitted
=== Sandbox denials ===
(bash) file-write-data /dev/tty
(bash) file-write-data /dev/ttys001
(bash) sysctl-read kern.ngroups
(bash) file-write-create /Users/nornagon/foo.txt
```
It operates by:
1. spawning `log stream` to watch system logs, and
2. tracking all descendant PIDs using kqueue + proc_listchildpids.
this is a "best-effort" technique, as `log stream` may drop logs(?), and
kqueue + proc_listchildpids isn't atomic and can end up missing very
short-lived processes. But it works well enough in my testing to be
useful :)
## Summary
- log and surface clipboard failures instead of silently ignoring them
when `Ctrl+V` pastes an image (`paste_image_to_temp_png()` now feeds an
error history cell)
- enable `arboard`’s `wayland-data-control` feature so native Wayland
sessions can deliver image data without XWayland
- keep the success path unchanged: valid images still attach and show
the `[image …]` placeholder as before
Fixes#4818
---------
Co-authored-by: Eric Traut <etraut@openai.com>
Co-authored-by: Jeremy Rose <172423086+nornagon-openai@users.noreply.github.com>
## Summary
- add a `hide_rate_limit_model_nudge` notice flag plus config edit
plumbing so the rate limit reminder preference is persisted and
documented
- extend the chat widget prompt with a "never show again" option, and
wire new app events so selecting it hides future nudges immediately and
writes the config
- add unit coverage and refresh the snapshot for the three-option prompt
## Testing
- `just fmt`
- `just fix -p codex-tui`
- `just fix -p codex-core`
- `cargo test -p codex-tui`
- `cargo test -p codex-core` *(fails at
`exec::tests::kill_child_process_group_kills_grandchildren_on_timeout`:
grandchild process still alive)*
------
[Codex
Task](https://chatgpt.com/codex/tasks/task_i_6910d7f407748321b2661fc355416994)
Add a `codex generate-json-schema` command for generating a JSON schema
bundle of app-server types, analogous to the existing `codex
generate-ts` command for Typescript.
The CLA action is designed to automatically lock a PR when it is closed.
This preserves the CLA agreement statements, preventing the contributor
from deleting them after the fact. However, this action is currently
locking PRs that are closed without merging. I'd like to keep such PRs
open so the contributor can respond with additional comments. I'm
currently manually unlocking PRs that I close, but I'd like to eliminate
this manual step.
I recently fixed a bug in [this
PR](https://github.com/openai/codex/pull/6285) that prevented Ctrl+C
from dismissing the login menu in the TUI and leaving the user unauthed.
A [user pointed out](https://github.com/openai/codex/issues/6418) that
this makes Ctrl+C can no longer be used to exit the app. This PR changes
the behavior so we exit the app rather than ignoring the Ctrl+C.
Update `codex generate-ts` to use the TS export code from
`app-server-protocol/src/export.rs`.
I realized there were two duplicate implementations of Typescript export
code:
- `app-server-protocol/src/export.rs`
- the `codex-protocol-ts` crate
The `codex-protocol-ts` crate that `codex generate-ts` uses is out of
date now since it doesn't handle the V2 namespace from:
https://github.com/openai/codex/pull/6212.
