obzerve 0.1.0__tar.gz

obzerve-0.1.0/.gitignore

@@ -0,0 +1,21 @@
+# Python
+.venv/
+__pycache__/
+*.py[cod]
+.pytest_cache/
+.ruff_cache/
+dist/
+*.egg-info/
+
+# Node
+node_modules/
+web/dist/
+
+# Local config
+.env
+.env.local
+
+# Editors / OS
+.idea/
+.vscode/
+.DS_Store

obzerve-0.1.0/CHANGELOG.md

@@ -0,0 +1,13 @@
+# Changelog
+
+All notable changes to the `obzerve` Python SDK. Versions follow [Semantic Versioning](https://semver.org/); while the version is 0.x, minor releases may change the API.
+
+## 0.1.0
+
+First release.
+
+- Tracing built on OpenTelemetry: `Observer.trace()`, `span()`, `@observe`, a bounded non-blocking export, and exact delivery counters (`Observer.stats()`).
+- Automatic instrumentation (`Observer.instrument()`) for `openai` (Chat Completions, Responses, Embeddings), `anthropic` (Messages, including streaming), `huggingface_hub` (`InferenceClient`) and `transformers` (`generate()` and text-generation pipelines): sync, async, and streams.
+- Replay: `@replayable` operations, automatic recordings of non-streaming instrumented calls, and `RECORDED` / `LLM_LIVE` / `LIVE` policies with divergence detection and side-effect guards.
+- Evaluations: `obzerve.evals.run_eval`, deterministic assertions, and the `obzerve` CLI (`case add`, `eval compare`, `eval gate`).
+- Payload redaction and truncation; W3C trace-context propagation across services.

obzerve-0.1.0/LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

obzerve-0.1.0/PKG-INFO

@@ -0,0 +1,170 @@
+Metadata-Version: 2.5
+Name: obzerve
+Version: 0.1.0
+Summary: Observability, replay, and regression testing for AI agents: OpenTelemetry tracing with automatic instrumentation for OpenAI, Anthropic, and Hugging Face models.
+Project-URL: Homepage, https://github.com/aryanfredericks/Obzerve
+Project-URL: Documentation, https://github.com/aryanfredericks/Obzerve/blob/main/sdk/python/README.md
+Project-URL: Source, https://github.com/aryanfredericks/Obzerve/tree/main/sdk/python
+Project-URL: Issues, https://github.com/aryanfredericks/Obzerve/issues
+Project-URL: Changelog, https://github.com/aryanfredericks/Obzerve/blob/main/sdk/python/CHANGELOG.md
+Author: Aryan Fredericks
+License-Expression: Apache-2.0
+License-File: LICENSE
+Keywords: agents,ai-agents,anthropic,evals,huggingface,llm,observability,openai,opentelemetry,replay,tracing,transformers
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Software Development :: Testing
+Classifier: Topic :: System :: Monitoring
+Classifier: Typing :: Typed
+Requires-Python: >=3.10
+Requires-Dist: opentelemetry-api>=1.27
+Requires-Dist: opentelemetry-exporter-otlp-proto-http>=1.27
+Requires-Dist: opentelemetry-sdk>=1.27
+Requires-Dist: requests>=2.28
+Provides-Extra: all
+Requires-Dist: anthropic>=2; extra == 'all'
+Requires-Dist: huggingface-hub>=2; extra == 'all'
+Requires-Dist: openai>=3; extra == 'all'
+Requires-Dist: torch; extra == 'all'
+Requires-Dist: transformers>=5; extra == 'all'
+Provides-Extra: anthropic
+Requires-Dist: anthropic>=2; extra == 'anthropic'
+Provides-Extra: huggingface
+Requires-Dist: huggingface-hub>=2; extra == 'huggingface'
+Provides-Extra: openai
+Requires-Dist: openai>=3; extra == 'openai'
+Provides-Extra: transformers
+Requires-Dist: torch; extra == 'transformers'
+Requires-Dist: transformers>=5; extra == 'transformers'
+Description-Content-Type: text/markdown
+
+# Obzerve Python SDK
+
+Observability, replay, and regression testing for AI agents.
+
+`obzerve` traces what your agent does: LLM calls, tool calls, retrieval, and handoffs. Traces go to a self-hosted [Obzerve](https://github.com/aryanfredericks/Obzerve) server, where you can see the timeline, latency, errors, token usage, and estimated cost of every request. When something goes wrong, the SDK can replay the exact execution from its recordings and turn it into a regression test that gates your next release.
+
+The SDK is built on OpenTelemetry and sends standard OTLP.
+
+```bash
+pip install obzerve                  # or: uv add obzerve
+pip install "obzerve[openai]"        # plus the client libraries you use: openai, anthropic,
+                                     # huggingface, transformers, or all
+```
+
+You need a running Obzerve server and a project API key. See the [getting-started guide](https://github.com/aryanfredericks/Obzerve/blob/main/docs/getting-started.md); `make stack-demo` starts one locally in containers.
+
+## Trace an agent in three lines
+
+```python
+from obzerve import Observer
+
+observer = Observer(endpoint="http://localhost:8000", api_key="obz_...", service_name="support-bot")
+observer.instrument()  # trace OpenAI, Anthropic and Hugging Face calls automatically
+
+
+def handle(ticket):
+    with observer.trace("support_bot", input={"ticket": ticket}) as trace:
+        answer = ...  # your agent: every LLM call inside becomes a span of this trace
+        trace.set_output(answer)
+        return answer
+```
+
+## Automatic instrumentation
+
+`observer.instrument()` patches the supported client libraries that are installed. Every model call becomes an LLM span inside the active trace, recording:
+- model and provider
+- sampling settings
+- input, output, and cached token counts
+- finish reason
+- inputs and outputs, subject to redaction and `capture_payloads`
+
+| Library | What is traced |
+| --- | --- |
+| `openai` | Chat Completions, Responses, and Embeddings: sync, async, and streams. OpenAI-compatible servers (vLLM, Ollama, TGI, LM Studio, ...) reached through the OpenAI client are recorded as `openai_compatible`, with their server address |
+| `anthropic` | `messages.create` (including `stream=True`) and the `messages.stream()` helper, sync and async. Cache reads and writes are included in input tokens, and cache reads are priced as cached tokens |
+| `huggingface_hub` | `InferenceClient.chat_completion` (also reached as `client.chat.completions.create`) and `text_generation`, sync, async, and streams. Covers the serverless Inference API, Inference Endpoints, and self-hosted TGI |
+| `transformers` | Any local model's `generate()` (model, device, prompt and generated token counts) and text-generation pipelines |
+
+```python
+observer.instrument(openai=True, anthropic=False)  # True: require it; False: skip it;
+# None (default): instrument it if installed
+handle = observer.instrument()
+handle.uninstrument()  # restore the original libraries
+```
+
+`transformers` loads PyTorch on import, so by default it is instrumented only if your code has already imported it. Pass `transformers=True` to instrument it regardless.
+
+Instrumentation never breaks your call: if a library changes shape, you lose telemetry, not the request. A streamed response stays open as one span until you finish reading it, close it, or leave its `with` block. Close streams you stop reading early.
+
+## Spans you define yourself
+
+```python
+from obzerve import SpanKind
+
+with observer.span("search_orders", kind=SpanKind.TOOL, input={"customer": cid}) as span:
+    orders = orders_api.search(cid)
+    span.set_output(orders)
+
+
+@observer.observe(kind=SpanKind.RETRIEVAL)  # decorator form, sync or async
+def retrieve(query: str) -> list[str]: ...
+```
+
+Exceptions mark the span as failed, with the error type and message. `span.record_retry(attempt, exc)` records retries. `span.set_llm(model=..., input_tokens=..., output_tokens=...)` records model usage for calls the instrumentation doesn't cover.
+
+## Replay and regression testing
+
+Calls through the instrumented libraries are recorded exactly: the whole response, rebuilt as the library's own response type during a replay. For your own tools, use `@observer.replayable`:
+
+```python
+@observer.replayable(kind=SpanKind.TOOL)
+def lookup_order(order_id: str) -> dict: ...
+
+
+@observer.replayable(kind=SpanKind.TOOL, side_effect=True)  # never re-run by a replay
+def refund(order_id: str) -> None: ...
+```
+
+Then reproduce a production request, without calling any model or tool:
+
+```python
+from obzerve.client import ObzerveClient
+from obzerve.replay import RECORDED, LLM_LIVE
+
+bundle = ObzerveClient("http://localhost:8000", api_key).replay_bundle(trace_id)
+with observer.replay(bundle, RECORDED):  # or LLM_LIVE: tools recorded, model live
+    handle(**bundle.root_input)
+```
+
+Turn the trace into a regression case, run your eval dataset against a change, and gate the release in CI:
+
+```bash
+obzerve case add <trace_id> --dataset regressions --name "refund flow" --mode recorded --no-error
+obzerve eval gate <baseline_run> <candidate_run> --max-regressions 0     # exits 1 on failure
+```
+
+Run datasets with `obzerve.evals.run_eval`. The [main README](https://github.com/aryanfredericks/Obzerve#replay-and-regression-testing) covers the full workflow.
+
+## Privacy and reliability defaults
+
+- **Redaction:** captured inputs and outputs are redacted (keys such as `api_key`, `password`, `authorization`, plus bearer and `sk-…` tokens) and truncated. `Observer(capture_payloads=False)` turns capture off entirely.
+- **Never in your way:** export runs in a background thread through a bounded buffer with short timeouts. A slow or unavailable server never blocks or fails your agent; telemetry is dropped instead. `observer.stats()` reports exactly how many spans were exported, failed, or dropped.
+- **No global takeover:** the SDK uses its own OpenTelemetry tracer provider and never replaces yours.
+- **Anthropic's own spans:** the Anthropic SDK can emit OpenTelemetry spans to your *global* provider. If you also send those to Obzerve, disable one of the two (`Anthropic(open_telemetry=False)`), or each call appears twice.
+
+## Current limits
+
+- Streamed responses are traced, but not yet recorded for replay. A replay that reaches one runs it live (`LLM_LIVE`) or fails with a clear error (`RECORDED`).
+- Replay matches calls by operation name and order. Concurrent calls of the same operation can't yet be told apart reliably.
+- Tested with Python 3.10–3.13, `openai` 3.x, `anthropic` 2.x, `huggingface_hub` 2.x, and `transformers` 5.x. Other versions are instrumented best-effort: an incompatible library logs a warning and is left alone.
+
+Licensed under [Apache-2.0](https://github.com/aryanfredericks/Obzerve/blob/main/LICENSE).

obzerve-0.1.0/README.md

@@ -0,0 +1,122 @@
+# Obzerve Python SDK
+
+Observability, replay, and regression testing for AI agents.
+
+`obzerve` traces what your agent does: LLM calls, tool calls, retrieval, and handoffs. Traces go to a self-hosted [Obzerve](https://github.com/aryanfredericks/Obzerve) server, where you can see the timeline, latency, errors, token usage, and estimated cost of every request. When something goes wrong, the SDK can replay the exact execution from its recordings and turn it into a regression test that gates your next release.
+
+The SDK is built on OpenTelemetry and sends standard OTLP.
+
+```bash
+pip install obzerve                  # or: uv add obzerve
+pip install "obzerve[openai]"        # plus the client libraries you use: openai, anthropic,
+                                     # huggingface, transformers, or all
+```
+
+You need a running Obzerve server and a project API key. See the [getting-started guide](https://github.com/aryanfredericks/Obzerve/blob/main/docs/getting-started.md); `make stack-demo` starts one locally in containers.
+
+## Trace an agent in three lines
+
+```python
+from obzerve import Observer
+
+observer = Observer(endpoint="http://localhost:8000", api_key="obz_...", service_name="support-bot")
+observer.instrument()  # trace OpenAI, Anthropic and Hugging Face calls automatically
+
+
+def handle(ticket):
+    with observer.trace("support_bot", input={"ticket": ticket}) as trace:
+        answer = ...  # your agent: every LLM call inside becomes a span of this trace
+        trace.set_output(answer)
+        return answer
+```
+
+## Automatic instrumentation
+
+`observer.instrument()` patches the supported client libraries that are installed. Every model call becomes an LLM span inside the active trace, recording:
+- model and provider
+- sampling settings
+- input, output, and cached token counts
+- finish reason
+- inputs and outputs, subject to redaction and `capture_payloads`
+
+| Library | What is traced |
+| --- | --- |
+| `openai` | Chat Completions, Responses, and Embeddings: sync, async, and streams. OpenAI-compatible servers (vLLM, Ollama, TGI, LM Studio, ...) reached through the OpenAI client are recorded as `openai_compatible`, with their server address |
+| `anthropic` | `messages.create` (including `stream=True`) and the `messages.stream()` helper, sync and async. Cache reads and writes are included in input tokens, and cache reads are priced as cached tokens |
+| `huggingface_hub` | `InferenceClient.chat_completion` (also reached as `client.chat.completions.create`) and `text_generation`, sync, async, and streams. Covers the serverless Inference API, Inference Endpoints, and self-hosted TGI |
+| `transformers` | Any local model's `generate()` (model, device, prompt and generated token counts) and text-generation pipelines |
+
+```python
+observer.instrument(openai=True, anthropic=False)  # True: require it; False: skip it;
+# None (default): instrument it if installed
+handle = observer.instrument()
+handle.uninstrument()  # restore the original libraries
+```
+
+`transformers` loads PyTorch on import, so by default it is instrumented only if your code has already imported it. Pass `transformers=True` to instrument it regardless.
+
+Instrumentation never breaks your call: if a library changes shape, you lose telemetry, not the request. A streamed response stays open as one span until you finish reading it, close it, or leave its `with` block. Close streams you stop reading early.
+
+## Spans you define yourself
+
+```python
+from obzerve import SpanKind
+
+with observer.span("search_orders", kind=SpanKind.TOOL, input={"customer": cid}) as span:
+    orders = orders_api.search(cid)
+    span.set_output(orders)
+
+
+@observer.observe(kind=SpanKind.RETRIEVAL)  # decorator form, sync or async
+def retrieve(query: str) -> list[str]: ...
+```
+
+Exceptions mark the span as failed, with the error type and message. `span.record_retry(attempt, exc)` records retries. `span.set_llm(model=..., input_tokens=..., output_tokens=...)` records model usage for calls the instrumentation doesn't cover.
+
+## Replay and regression testing
+
+Calls through the instrumented libraries are recorded exactly: the whole response, rebuilt as the library's own response type during a replay. For your own tools, use `@observer.replayable`:
+
+```python
+@observer.replayable(kind=SpanKind.TOOL)
+def lookup_order(order_id: str) -> dict: ...
+
+
+@observer.replayable(kind=SpanKind.TOOL, side_effect=True)  # never re-run by a replay
+def refund(order_id: str) -> None: ...
+```
+
+Then reproduce a production request, without calling any model or tool:
+
+```python
+from obzerve.client import ObzerveClient
+from obzerve.replay import RECORDED, LLM_LIVE
+
+bundle = ObzerveClient("http://localhost:8000", api_key).replay_bundle(trace_id)
+with observer.replay(bundle, RECORDED):  # or LLM_LIVE: tools recorded, model live
+    handle(**bundle.root_input)
+```
+
+Turn the trace into a regression case, run your eval dataset against a change, and gate the release in CI:
+
+```bash
+obzerve case add <trace_id> --dataset regressions --name "refund flow" --mode recorded --no-error
+obzerve eval gate <baseline_run> <candidate_run> --max-regressions 0     # exits 1 on failure
+```
+
+Run datasets with `obzerve.evals.run_eval`. The [main README](https://github.com/aryanfredericks/Obzerve#replay-and-regression-testing) covers the full workflow.
+
+## Privacy and reliability defaults
+
+- **Redaction:** captured inputs and outputs are redacted (keys such as `api_key`, `password`, `authorization`, plus bearer and `sk-…` tokens) and truncated. `Observer(capture_payloads=False)` turns capture off entirely.
+- **Never in your way:** export runs in a background thread through a bounded buffer with short timeouts. A slow or unavailable server never blocks or fails your agent; telemetry is dropped instead. `observer.stats()` reports exactly how many spans were exported, failed, or dropped.
+- **No global takeover:** the SDK uses its own OpenTelemetry tracer provider and never replaces yours.
+- **Anthropic's own spans:** the Anthropic SDK can emit OpenTelemetry spans to your *global* provider. If you also send those to Obzerve, disable one of the two (`Anthropic(open_telemetry=False)`), or each call appears twice.
+
+## Current limits
+
+- Streamed responses are traced, but not yet recorded for replay. A replay that reaches one runs it live (`LLM_LIVE`) or fails with a clear error (`RECORDED`).
+- Replay matches calls by operation name and order. Concurrent calls of the same operation can't yet be told apart reliably.
+- Tested with Python 3.10–3.13, `openai` 3.x, `anthropic` 2.x, `huggingface_hub` 2.x, and `transformers` 5.x. Other versions are instrumented best-effort: an incompatible library logs a warning and is left alone.
+
+Licensed under [Apache-2.0](https://github.com/aryanfredericks/Obzerve/blob/main/LICENSE).

obzerve-0.1.0/pyproject.toml

@@ -0,0 +1,72 @@
+[project]
+name = "obzerve"
+dynamic = ["version"]
+description = "Observability, replay, and regression testing for AI agents: OpenTelemetry tracing with automatic instrumentation for OpenAI, Anthropic, and Hugging Face models."
+readme = "README.md"
+license = "Apache-2.0"
+license-files = ["LICENSE"]
+authors = [{ name = "Aryan Fredericks" }]
+requires-python = ">=3.10"
+keywords = [
+    "observability", "tracing", "opentelemetry", "llm", "agents", "ai-agents",
+    "openai", "anthropic", "huggingface", "transformers", "replay", "evals",
+]
+classifiers = [
+    "Development Status :: 3 - Alpha",
+    "Intended Audience :: Developers",
+    "Operating System :: OS Independent",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3 :: Only",
+    "Programming Language :: Python :: 3.10",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "Topic :: Software Development :: Testing",
+    "Topic :: System :: Monitoring",
+    "Topic :: Scientific/Engineering :: Artificial Intelligence",
+    "Typing :: Typed",
+]
+dependencies = [
+    "opentelemetry-api>=1.27",
+    "opentelemetry-sdk>=1.27",
+    "opentelemetry-exporter-otlp-proto-http>=1.27",
+    "requests>=2.28",
+]
+
+[project.optional-dependencies]
+# Convenience pins for the client libraries Obzerve instruments. The core package
+# never imports them; instrumentation is best-effort for other versions.
+openai = ["openai>=3"]
+anthropic = ["anthropic>=2"]
+huggingface = ["huggingface_hub>=2"]
+transformers = ["transformers>=5", "torch"]
+all = ["obzerve[openai,anthropic,huggingface,transformers]"]
+
+[project.urls]
+Homepage = "https://github.com/aryanfredericks/Obzerve"
+Documentation = "https://github.com/aryanfredericks/Obzerve/blob/main/sdk/python/README.md"
+Source = "https://github.com/aryanfredericks/Obzerve/tree/main/sdk/python"
+Issues = "https://github.com/aryanfredericks/Obzerve/issues"
+Changelog = "https://github.com/aryanfredericks/Obzerve/blob/main/sdk/python/CHANGELOG.md"
+
+[project.scripts]
+obzerve = "obzerve.cli:main"
+
+[build-system]
+requires = ["hatchling>=1.27"]
+build-backend = "hatchling.build"
+
+[tool.hatch.version]
+path = "src/obzerve/_version.py"
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/obzerve"]
+
+[tool.hatch.build.targets.sdist]
+include = ["src/obzerve", "tests", "README.md", "CHANGELOG.md", "LICENSE"]
+
+[tool.ruff]
+# The SDK supports older Pythons than the rest of the workspace: lint it as 3.10 code
+# so pyupgrade never suggests 3.11+ features (e.g. datetime.UTC).
+extend = "../../pyproject.toml"
+target-version = "py310"

obzerve-0.1.0/src/obzerve/__init__.py

@@ -0,0 +1,7 @@
+"""Obzerve Python SDK: trace AI-agent workflows with OpenTelemetry."""
+
+from obzerve.attributes import SpanKind
+from obzerve.delivery import DeliveryStats
+from obzerve.observer import Observer, ObzerveSpan, __version__
+
+__all__ = ["DeliveryStats", "Observer", "ObzerveSpan", "SpanKind", "__version__"]

obzerve-0.1.0/src/obzerve/_version.py

@@ -0,0 +1,3 @@
+"""Obzerve SDK version (single source of truth for the package metadata)."""
+
+__version__ = "0.1.0"