Hi Deep Agents community! 👋
Deep Agents follows the LangChain OSS release and versioning policies documented here:
You can find package releases on the Deep Agents GitHub Releases page.
Package status
Deep Agents is currently a pre-1.0 package under active development.
That means:
- We try to minimize breaking changes where possible.
- APIs may still evolve between minor versions as the project matures.
- Patch releases are intended for bug fixes and minor improvements without breaking changes.
- After Deep Agents reaches
1.0, it will follow the same semantic versioning and LTS policies as LangChain and LangGraph.
For more detail, see the Deep Agents section of the release policy.
Stable releases
What: Stable releases are the default releases published for normal use.
Version format: 0.x.y while Deep Agents is pre-1.0.
Release cadence: Released as needed. Patch releases may happen frequently for bug fixes, security updates, and minor improvements.
Stability: Recommended for most users, with the caveat that Deep Agents is still pre-1.0. Minor version upgrades may include API changes.
How to install or upgrade:
pip install -U deepagents
To pin a specific version:
pip install deepagents==0.6.11
Pre-releases
What: Pre-releases are early previews of upcoming changes. These may include alpha, beta, or release-candidate builds.
Examples:
0.7.0a1 — alpha
0.7.0b1 — beta
0.7.0rc1 — release candidate
Stability: Use with caution. Pre-releases may contain regressions, incomplete features, or APIs that change before the stable release.
How to install:
pip install --pre deepagents
Or install a specific pre-release:
pip install deepagents==0.7.0a1
If you use a pre-release, please pin the exact version and test carefully before using it in production.
Changelog and release notes
There are two useful places to follow changes:
Breaking changes
While Deep Agents is pre-1.0:
- Minor releases, such as
0.6.0 → 0.7.0, may include breaking changes.
- Patch releases, such as
0.6.10 → 0.6.11, are intended to be backward-compatible bug fixes and minor improvements.
- Migration guidance will be included in release notes when significant changes occur.
After Deep Agents reaches 1.0, breaking changes will be reserved for major releases.
Security reports
Please do not report security vulnerabilities in public GitHub issues.
To report a vulnerability:
- Submit a security advisory through the Security tab of the relevant GitHub repository.
- Email
security@langchain.dev to notify us that you filed a report and identify the repository.
See the full LangChain OSS security policy for details.
FAQ
Where can I find the latest Deep Agents release?
Use the Deep Agents GitHub Releases page.
Where can I find notable release notes?
Use both the LangChain OSS Python changelog and the Deep Agents GitHub Releases page. The changelog highlights notable user-facing Deep Agents releases, while GitHub Releases has the package-level release history, including patch releases and release assets.
Is Deep Agents production-ready?
Deep Agents is actively used and developed, but it is still pre-1.0. We recommend pinning versions for production deployments and reviewing release notes before upgrading minor versions.
Should I use pre-releases?
Use pre-releases only if you want early access to upcoming changes and are comfortable testing against possible regressions. Please report issues you find on the Deep Agents issue tracker.
Hi Deep Agents community! 👋
Deep Agents follows the LangChain OSS release and versioning policies documented here:
You can find package releases on the Deep Agents GitHub Releases page.
Package status
Deep Agents is currently a pre-1.0 package under active development.
That means:
1.0, it will follow the same semantic versioning and LTS policies as LangChain and LangGraph.For more detail, see the Deep Agents section of the release policy.
Stable releases
What: Stable releases are the default releases published for normal use.
Version format:
0.x.ywhile Deep Agents is pre-1.0.Release cadence: Released as needed. Patch releases may happen frequently for bug fixes, security updates, and minor improvements.
Stability: Recommended for most users, with the caveat that Deep Agents is still pre-1.0. Minor version upgrades may include API changes.
How to install or upgrade:
To pin a specific version:
Pre-releases
What: Pre-releases are early previews of upcoming changes. These may include alpha, beta, or release-candidate builds.
Examples:
0.7.0a1— alpha0.7.0b1— beta0.7.0rc1— release candidateStability: Use with caution. Pre-releases may contain regressions, incomplete features, or APIs that change before the stable release.
How to install:
Or install a specific pre-release:
If you use a pre-release, please pin the exact version and test carefully before using it in production.
Changelog and release notes
There are two useful places to follow changes:
Breaking changes
While Deep Agents is pre-1.0:
0.6.0→0.7.0, may include breaking changes.0.6.10→0.6.11, are intended to be backward-compatible bug fixes and minor improvements.After Deep Agents reaches
1.0, breaking changes will be reserved for major releases.Security reports
Please do not report security vulnerabilities in public GitHub issues.
To report a vulnerability:
security@langchain.devto notify us that you filed a report and identify the repository.See the full LangChain OSS security policy for details.
FAQ
Where can I find the latest Deep Agents release?
Use the Deep Agents GitHub Releases page.
Where can I find notable release notes?
Use both the LangChain OSS Python changelog and the Deep Agents GitHub Releases page. The changelog highlights notable user-facing Deep Agents releases, while GitHub Releases has the package-level release history, including patch releases and release assets.
Is Deep Agents production-ready?
Deep Agents is actively used and developed, but it is still pre-1.0. We recommend pinning versions for production deployments and reviewing release notes before upgrading minor versions.
Should I use pre-releases?
Use pre-releases only if you want early access to upcoming changes and are comfortable testing against possible regressions. Please report issues you find on the Deep Agents issue tracker.