releases
Coverage, reference pages, tools, and guides connected to this topic.
-
e2b @e2b/[email protected] released
e2b/[email protected] fixes `e2b auth login` on headless machines by avoiding the `xdg-open` crash path and handling missing browser openers synchronously. If interactive login still isn’t possible, it now prints the login URL for manual use and points users to `E2B_API_KEY` as a fallback.
-
langgraph sdk==0.3.15 released
LangGraph SDK 0.3.15 adds metadata filtering to cron search/count, making it easier for agent builders to target and manage scheduled jobs by metadata, and it bumps core packages including `langchain-core` to 1.4.0, `langgraph` to 1.2.1, and promotes alpha packages to official versions. It also fixes a concrete SDK bug by percent-encoding caller-supplied identifiers in URL paths, which should prevent malformed requests and improve reliability when IDs contain special characters.
-
pydantic-ai 1.102.0 released
Pydantic AI 1.102.0 fixes a security issue in URL validation by expanding IPv6 transition-form handling to block an SSRF cloud-metadata bypass, but it only matters if you explicitly use `FileUrl(force_download='allow-local')` with untrusted URLs on NAT64/ISATAP networks. It also adds support for additional IPv6 transition forms in URL validation, which makes agent file-download and URL-handling flows more robust against edge-case network inputs.
-
pydantic-ai 1.100.0 released
Pydantic AI 1.100.0 is mainly a security release for agent builders: it normalizes IPv6 transition forms in URL validation to close an SSRF cloud-metadata blocklist bypass affecting `FileUrl(..., force_download='allow-local')` when the URL can be influenced by untrusted input. It does not appear to add major new agent capabilities or performance improvements in the excerpt, and the fix was already shipped in v1.99.0.
-
crewai 1.14.5 released
CrewAI 1.14.5 makes a breaking executor shift by deprecating `CrewAgentExecutor` in favor of `AgentExecutor`, while also renaming `EXASearchTool` to `ExaSearchTool` and fixing the status endpoint path to `/status/{kickoff_id}`. It adds a new `restore_from_state_id` kickoff option, improves Daytona sandbox tools, and tightens runtime behavior with fixes for memory leaks, streamed tool calls, async batch output preservation, and more reliable final-answer handling.
-
pydantic-ai 1.99.0 released
Pydantic AI 1.99.0 adds support for the `gemini-3.5-flash` model, giving agent builders another model option to use in production workflows. It also fixes an OpenAI strict-schema bug involving regex lookarounds, which should improve reliability when validating or generating structured outputs.
-
e2b [email protected] released
[email protected] adds abort support across the JS and Python SDKs, letting AI agent builders pass an `AbortSignal`/per-call API options to cancel in-flight requests for sandbox creation/connection, command execution, file and volume operations, and paginator requests. This is a concrete reliability and control upgrade: paginator calls can now override connection-level options per request, and aborted requests fail with an `AbortError` instead of continuing to run.
-
pydantic-ai 1.98.0 released
pydantic-ai 1.98.0 adds OpenAI Responses input token counting via `OpenAIResponsesModel.count_tokens`, which helps agent builders estimate usage/costs more accurately before sending requests. It also replaces the old `Agent(tool_retries=..., output_retries=...)` knobs with a single `retries: int | AgentRetries` setting, and deprecates `pydantic_ai.ext.aci` (`tool_from_aci` and `ACIToolset`) as part of V2 preparation.
-
e2b [email protected] released
E2B `[email protected]` adds an optional `name` parameter to `createSnapshot` and now returns snapshot names, making snapshots easier to identify and manage programmatically. There are no breaking changes or major performance updates in this patch, just a small SDK capability improvement for AI agent builders.
-
letta 0.16.8 released
Lepta 0.16.8 is a small patch release with no new agent-builder features or breaking API changes; the main functional fix is an update to workflows. It also hardens sandbox-to-server tool result transport by switching from pickle to JSON, which is a security improvement rather than a performance change.
-
pydantic-ai 1.97.0 released
Pydantic AI 1.97.0 adds `OnlineEvaluator.run_on_errors` so evals can now run on failed calls, and it changes Google provider setup by splitting `GoogleProvider(vertexai=True|False)` into `GoogleProvider` plus `GoogleCloudProvider` while renaming provider IDs from `google-gla:` to `google:` and `google-vertex:` to `google-cloud:`. It also introduces `MCPToolset` based on `fastmcp-slim[client]` and deprecates `MCPServer*` and `FastMCPToolset`, while setting `ModelResponse.state` to `incomplete` during streaming so builders can more accurately track in-progress responses.
-
pydantic-ai 1.96.0 released
Pydantic-ai 1.96.0 introduces V2 preparation changes, including the new `openai-chat:` prefix with warnings for bare `openai:` (which will switch to Responses API in v2), deprecation of `Agent(history_processors=)` in favor of `capabilities=[ProcessHistory(...)]`, and deprecation of method-style accessors like `result.usage()` for property-style equivalents. It also deprecates `AGUIApp`, `Agent.to_ag_ui()`, and related shims in favor of `AGUIAdapter`, signaling upcoming shifts for AI agent builders toward more standardized APIs.
-
pydantic-ai 1.95.0 released
Pydantic AI 1.95.0 introduces native Tool Search for Anthropic and OpenAI with custom strategies on any provider, an Instrumentation capability replacing Agent(instrument=...), and structured output plus tool combination support for Gemini 3. It prepares for V2 by renaming “built-in tools” to “native tools”, deprecating old fields, and registering them via capabilities=[NativeTool(...)].
-
langgraph 1.2.0 released
LangGraph 1.2.0 introduces durable error-handler resume across host crashes, enabling agents to recover from infrastructure failures without losing state, and adds `set_node_defaults()` to StateGraph for simplified node configuration. The release also optimizes checkpoint management with delta channel snapshots after max supersteps, improving performance for long-running agent workflows.
-
pydantic-ai 1.94.0 released
Pydantic-ai 1.94.0 adds support for OpenAI's multiple system messages capability through a new profile flag, enabling more flexible prompt structuring for OpenAI models. The release also removes `mistralai` as a direct dependency, streamlining the package's dependency footprint.
-
e2b [email protected] released
E2B version 2.20.0 introduces minor compatibility improvements for Turbopack, enabling smoother integration with modern bundling tools for AI agent development. No breaking changes or major new capabilities are noted, with the update focusing solely on this targeted enhancement.
-
pydantic-ai 1.93.0 released
Pydantic AI 1.93.0 adds a `tool_choice` setting for more control over model tool selection and improves event handling by yielding `OutputToolCallEvent`/`OutputToolResultEvent` for output tool calls while deprecating function-tool events for failing cases. The release also fixes a bug where spawned tasks weren't properly drained during agent cancellation, improving reliability in concurrent scenarios.
-
pydantic-ai 1.92.0 released
Pydantic AI 1.92.0 introduces Anthropic task budget support and runtime `output_retries` override with deprecation of the old `retries` field, enhancing control over AI agent execution and reliability. It also fixes key bugs like streaming response cleanup on cancellation, MCP session task isolation to prevent exit scope errors, and proper population of `RunContext` with run/conversation IDs and metadata.
-
autogen python-v0.7.5 released
AutoGen v0.7.5 adds linear memory support in RedisMemory, enabling more scalable and efficient long‑running agent conversations. It also introduces thinking mode for the Anthropic client and fixes several streaming, tool‑call, and correlation issues that improve reliability and performance for agent builders.
-
crewai 1.14.4 released
CrewAI 1.14.4 introduces enhanced cloud provider support with custom persistence keys for @persist, Responses API for Azure OpenAI, and new search/research tools via Tavily and You.com MCP integration. The release also includes critical bug fixes for JSON parsing, tool call preservation, and multimodal input handling, improving reliability for production agent deployments.
-
langgraph sdk==0.3.14 released
LangGraph SDK 0.3.14 introduces a `return_minimal` parameter for threads update operations, enabling more efficient API responses for AI agent builders. The release also includes streaming transformer infrastructure and support for `stream_events(version='v3')` on Pregel, providing enhanced control over event streaming in agent workflows.
-
letta 0.16.7 released
Letta 0.16.7 raises the default global context window from 32k to 128k and fixes the context window reset bug, with a completely overhauled compaction system that eliminates most manual configuration workarounds for self-hosted users. Block limits are no longer enforced, allowing blocks to grow freely, though users must now manage block size through alternative means if they were previously relying on limits to control per-turn costs.