Releases: launchdarkly/python-server-sdk-ai
Releases · launchdarkly/python-server-sdk-ai
launchdarkly-server-sdk-ai-openai: v0.5.0
Immutable
release. Only release title and notes can be modified.
0.5.0 (2026-05-05)
Features
- Add judge evaluation support to agent graphs (#142) (3d5a6a9)
- Support conversation history directly in AI Provider model runners (#166) (4bb3e78)
- Update OpenAI graph runner to return AgentGraphRunnerResult with GraphMetrics (#155) (388b7af)
- Update OpenAI runners to implement Runner protocol returning RunnerResult (#149) (382e662)
Bug Fixes
launchdarkly-server-sdk-ai-langchain: v0.6.0
Immutable
release. Only release title and notes can be modified.
0.6.0 (2026-05-05)
Features
- Add judge evaluation support to agent graphs (#142) (3d5a6a9)
- Migrate LangGraph runner to AgentGraphRunnerResult; clean up legacy shape detection (#156) (efa8e00)
- Support conversation history directly in AI Provider model runners (#166) (4bb3e78)
- Update LangChain runners to implement Runner protocol returning RunnerResult (#150) (62a8e25)
Bug Fixes
launchdarkly-server-sdk-ai: v0.19.0
Immutable
release. Only release title and notes can be modified.
0.19.0 (2026-05-05)
⚠ BREAKING CHANGES
- StructuredResponse replaced by RunnerResult with new "parsed" property
- AgentResult replaced by RunnerResult and Managed Result
- Removed ModelRunner and AgentRunner protocols
- Removed invoke_method, invoke_structured_model from AIProvider base class.
- ModelResponse was replaced by RunnerResult
- Add ManagedResult, RunnerResult, and Runner protocol; rename invoke() to run() (#148)
- Swap track_metrics_of parameter order to match spec (#144)
Features
- Add evaluations support to ManagedAgent.run() (#153) (442f46a)
- Add judge evaluation support to agent graphs (#142) (3d5a6a9)
- Add ManagedGraphResult, GraphMetricSummary, and AgentGraphRunnerResult types (#151) (301e24c)
- Add ManagedResult, RunnerResult, and Runner protocol; rename invoke() to run() (#148) (88d4ddc)
- Add root-level tools map with customParameters to AI Config types (#141) (f17c535)
- bake sampling_rate into Judge at construction; simplify Evaluator to List[Judge] (#159) (86c79e6)
- Update LangChain runners to implement Runner protocol returning RunnerResult (#150) (62a8e25)
Bug Fixes
- Add runtime DeprecationWarnings to deprecated methods (#145) (2189b81)
- AgentResult replaced by RunnerResult and Managed Result (fbb0b4b)
- build judge input as string; strip legacy judge config messages (#165) (e6942a6)
- Fall back to model.parameters.tools when root tools absent (#146) (2c30d75)
- Graph tracking refactor — ManagedAgentGraph drives tracking for new runner shape (#154) (20a5020)
- ModelResponse was replaced by RunnerResult (fbb0b4b)
- parse model.parameters.tools as list (#160) (fb53e99)
- reference correct PyPI package names in provider load error messages (#164) (48761c9)
- Removed invoke_method, invoke_structured_model from AIProvider base class. (fbb0b4b)
- Removed ModelRunner and AgentRunner protocols (fbb0b4b)
- Replace done_callback with coroutine chain for judge tracking (#147) (1e1f36b)
- StructuredResponse replaced by RunnerResult with new "parsed" property (fbb0b4b)
- Swap track_metrics_of parameter order to match spec (#144) (53db736)
launchdarkly-server-sdk-ai-openai: v0.4.0
Immutable
release. Only release title and notes can be modified.
launchdarkly-server-sdk-ai-langchain: v0.5.0
Immutable
release. Only release title and notes can be modified.
launchdarkly-server-sdk-ai: v0.18.0
Immutable
release. Only release title and notes can be modified.
0.18.0 (2026-04-21)
⚠ BREAKING CHANGES
- Add per-execution runId, at-most-once tracking, and cross-process tracker resumption (#133)
- rename track_latency to track_duration on AIGraphTracker (#138)
- Move graph_key to AIConfigTracker instantiation (#134)
- Flatten JudgeResponse and EvalScore into new JudgeResult (#132)
Features
- Add per-execution runId, at-most-once tracking, and cross-process tracker resumption (#133) (68685cd)
- Flatten JudgeResponse and EvalScore into new JudgeResult (#132) (af4e463)
- Move graph_key to AIConfigTracker instantiation (#134) (20fff24)
- rename track_latency to track_duration on AIGraphTracker (#138) (05758a7)
launchdarkly-server-sdk-ai-langchain: v0.4.1
Immutable
release. Only release title and notes can be modified.
launchdarkly-server-sdk-ai-optimization: v0.1.0
Immutable
release. Only release title and notes can be modified.
0.1.0 (2026-04-02)
⚠ BREAKING CHANGES
- Bump minimum LangChain version to 1.0.0
Features
- Add optimization package stub (#109) (ebd5166)
- Adds optimization package stub (58b7731)
- Adds optimization package stub (cc85a05)
- Bump minimum LangChain version to 1.0.0 (dc592c5)
- Drop support for python 3.9 (#114) (dc592c5)
Bug Fixes
- consistency with other makefiles (b9a5601)
launchdarkly-server-sdk-ai-openai: v0.3.0
Immutable
release. Only release title and notes can be modified.
0.3.0 (2026-04-02)
⚠ BREAKING CHANGES
- Bump minimum LangChain version to 1.0.0
- Extract shared utilities to openai_helper
- Extract shared utilities to langchain_helper
- Restructure provider factory and support additional create methods (#102)
Features
- Add LangGraphAgentGraphRunner (56ce0fd)
- Add LDAIClient.create_agent() returning ManagedAgent (53fd95e)
- Add ManagedAgentGraph support (#111) (56ce0fd)
- Add ModelRunner ABC with invoke_model() and (453c71c)
- Add OpenAIAgentGraphRunner (56ce0fd)
- Add OpenAIAgentRunner with agentic tool-calling loop (53fd95e)
- add optimization package stub (872e81e)
- Adds optimization package stub (58b7731)
- Bump minimum LangChain version to 1.0.0 (dc592c5)
- Deprecated Chat object in favor of ManagedModel (453c71c)
- Deprecated create_chat(), use create_model() on the LDAIClient (453c71c)
- Drop support for python 3.9 (#114) (dc592c5)
- Extract shared utilities to langchain_helper (453c71c)
- Extract shared utilities to openai_helper (453c71c)
- Introduce ManagedAgent and AgentRunner implementations (#110) (53fd95e)
- Introduce ManagedModel and ModelRunner (PR-3) (#104) (453c71c)
- Restructure provider factory and support additional create methods (#102) (e6e4907)
launchdarkly-server-sdk-ai-langchain: v0.4.0
Immutable
release. Only release title and notes can be modified.
0.4.0 (2026-04-02)
⚠ BREAKING CHANGES
- Bump minimum LangChain version to 1.0.0
- Split track_metrics_of into sync and async variants (#112)
- Remove node-scoped methods from AIGraphTracker (track_node_invocation, track_tool_call, track_node_judge_response), use related AIConfigTracker methods instead
- Extract shared utilities to openai_helper
- Extract shared utilities to langchain_helper
- Restructure provider factory and support additional create methods (#102)
Features
- Add get_ai_usage_from_response to openai_helper (4fab18f)
- Add get_tool_calls_from_response and sum_token_usage_from_messages to langchain_helper (4fab18f)
- Add LangGraphAgentGraphRunner (56ce0fd)
- Add LDAIClient.create_agent() returning ManagedAgent (53fd95e)
- Add ManagedAgentGraph support (#111) (56ce0fd)
- Add ModelRunner ABC with invoke_model() and (453c71c)
- Add OpenAIAgentGraphRunner (56ce0fd)
- Add OpenAIAgentRunner with agentic tool-calling loop (53fd95e)
- add optimization package stub (872e81e)
- Add optional graph_key to all LDAIConfigTracker track_* methods for graph correlation (4fab18f)
- Add track_tool_call/track_tool_calls to LDAIConfigTracker (4fab18f)
- Adds optimization package stub (58b7731)
- Bump minimum LangChain version to 1.0.0 (dc592c5)
- Deprecated Chat object in favor of ManagedModel (453c71c)
- Deprecated create_chat(), use create_model() on the LDAIClient (453c71c)
- Drop support for python 3.9 (#114) (dc592c5)
- Extract shared utilities to langchain_helper (453c71c)
- Extract shared utilities to openai_helper (453c71c)
- Introduce ManagedAgent and AgentRunner implementations (#110) (53fd95e)
- Introduce ManagedModel and ModelRunner (PR-3) (#104) (453c71c)
- Restructure provider factory and support additional create methods (#102) (e6e4907)
- Split track_metrics_of into sync and async variants (#112) (4fab18f)
Bug Fixes
- make AIGraphTracker.track_total_tokens accept Optional[TokenUsage], skip when None or total <= 0 (4fab18f)
- Remove node-scoped methods from AIGraphTracker (track_node_invocation, track_tool_call, track_node_judge_response), use related AIConfigTracker methods instead (4fab18f)
- use time.perf_counter_ns() for sub-millisecond precision in duration calculations (4fab18f)