release: 0.34.0

[stainless-app]

Automated Release PR

0.34.0 (2026-03-06)

Full Changelog: v0.33.0...v0.34.0

Features

• api: api update (0f66c9d)
• api: api update (b76b873)
• api: api update (b75e2b3)
• api: api update (d330dea)
• api: api update (e01d911)

Chores

• api: python trusted publisher (3d7a283)
• internal: add request options to SSE classes (6b5c9ad)
• internal: make test_proxy_environment_variables more resilient (3cc3332)
• internal: make test_proxy_environment_variables more resilient to env (2c99a7f)

---

This pull request is managed by Stainless's GitHub App.

The semver version number is based on included commit messages. Alternatively, you can manually set the version number in the title of this pull request.

For a better experience, it is recommended to use either rebase-merge or squash-merge when merging this pull request.

🔗 Stainless website
📚 Read the docs
🙋 Reach out for help or questions

[stainless-app]

🧪 Testing

To try out this version of the SDK:

pip install 'https://pkg.stainless.com/s/hyperspell-python/056466258f6dff9fe7709a85f7f5879e31b332a7/hyperspell-0.33.0-py3-none-any.whl'

Expires at: Sun, 05 Apr 2026 01:29:04 GMT
Updated at: Fri, 06 Mar 2026 01:29:04 GMT

[canaries-inc]

🐤 Canary Summary

This release expands the Hyperspell Python SDK to support two new data source integrations:

• GitHub integration added as a new memory source for indexing and searching code repositories
• Trace integration added as a new memory source for indexing and searching trace data
• New AI models (gpt-oss-20b, gpt-oss-120b) added for answer generation in memory search
• Resource ID filtering capability added to memory search for scoping queries to specific documents
• PyPI publishing workflow migrated to OIDC trusted publisher authentication
• Internal SDK improvements for request options handling in streaming responses

Note: This is a Python SDK library, not a UI application. The type changes enable new integration options that would appear in UI components of applications consuming this SDK (e.g., integration selection dropdowns, connection management interfaces, search filters).

---

[canaries-inc]

🐤 Canary Proposed Tests

No testable user journeys found for this PR.

[entelligence-ai-pr-reviews]

Correctness: Removing the PYPI_TOKEN validation lets the script pass even when the token is unset, which can cause release failures later (or worse, publish without auth). Reintroduce the check to fail fast when the secret is missing. 🚨

🤖 AI Agent Prompt for Cursor/Windsurf

📋 Copy this prompt to your AI coding assistant (Cursor, Windsurf, etc.) to get help fixing this issue

In `bin/check-release-environment` around lines 2-4, reintroduce the PYPI_TOKEN validation block so the script errors when the token is missing. Restore the removed `if [ -z "${PYPI_TOKEN}" ]; then ... fi` block before `lenErrors`.

[entelligence-ai-pr-reviews]

Duplicate Code: ⚠️ Duplicate Code Detected (Similarity: 95%)

This function Stream.__init__ duplicates existing code.

📍 Original Location:

src/hyperspell/_streaming.py:121-142

Function: AsyncStream.__init__

💡 Recommendation:
Introduce an abstract base class (e.g., BaseStream) that holds the shared state: _options, _cast_to, _client, _timeout, and their initialization. Both Stream and AsyncStream can inherit from it, eliminating the duplicated __init__ initialization block. Alternatively, if the sync/async split makes inheritance awkward, at least extract the shared options initialization into a _init_options(self, options) helper or use a @dataclass-style mixin.

Consider importing and reusing the existing function instead of duplicating the logic.

[entelligence-ai-pr-reviews]

Duplicate Code: ⚠️ Duplicate Code Detected (Similarity: 97%)

This function TestClient.test_proxy_environment_variables duplicates existing code.

📍 Original Location:

tests/test_client.py:1998-2006 (async version)

Function: TestAsyncClient.test_proxy_environment_variables

💡 Recommendation:
Extract the proxy env var cleanup into a pytest fixture or a shared helper function (e.g., clear_proxy_env(monkeypatch)) and call it at the start of both test_proxy_environment_variables methods. Alternatively, a @pytest.fixture(autouse=True) scoped to the proxy test class could handle cleanup transparently.

Consider importing and reusing the existing function instead of duplicating the logic.