@letta-ai/letta-client 1.7.12 → 1.8.1

package/CHANGELOG.md

@@ -1,5 +1,56 @@
 # Changelog
 
+## 1.8.1 (2026-03-24)
+
+Full Changelog: [v1.8.0...v1.8.1](https://github.com/letta-ai/letta-node/compare/v1.8.0...v1.8.1)
+
+### Features
+
+* add idempotency for conversation streaming ([f640d1d](https://github.com/letta-ai/letta-node/commit/f640d1d810376cedb8b6709c6d2b5d16b615eecc))
+* **core:** add created_by_id and last_updated_by_id filters to list/count agents ([d090311](https://github.com/letta-ai/letta-node/commit/d090311ec93e9566c8109d6f3cdb5a47e3206d71))
+* **core:** sort conversations by last_message_at ([ea25296](https://github.com/letta-ai/letta-node/commit/ea25296a2d12aafcf30944113e26265b0f085af5))
+
+
+### Reverts
+
+* undo PR [#10058](https://github.com/letta-ai/letta-node/issues/10058) system-prompt warning changes ([12138c6](https://github.com/letta-ai/letta-node/commit/12138c6bb72d0ccb58c149b80c1eed152ec7af27))
+
+
+### Chores
+
+* **ci:** skip lint on metadata-only changes ([a7488f5](https://github.com/letta-ai/letta-node/commit/a7488f5163c41df6d76b226ededc520be1e5d778))
+* **internal:** update gitignore ([ac6dff3](https://github.com/letta-ai/letta-node/commit/ac6dff340f4f0229c05e174351957a47459a8b00))
+
+## 1.8.0 (2026-03-20)
+
+Full Changelog: [v1.7.12...v1.8.0](https://github.com/letta-ai/letta-node/compare/v1.7.12...v1.8.0)
+
+### Features
+
+* add zai coding provider ([88e56a9](https://github.com/letta-ai/letta-node/commit/88e56a9498385d6a37f2c96806c882a856cd6f68))
+* always create otid for request ([488b66b](https://github.com/letta-ai/letta-node/commit/488b66b6443a290d49e84a2d25d11b4c0b007822))
+
+
+### Bug Fixes
+
+* surface error when system prompt is large during compaction ([be4ef7f](https://github.com/letta-ai/letta-node/commit/be4ef7f6d547247430cda5f72e7bce1939896c27))
+
+
+### Chores
+
+* **internal:** tweak CI branches ([176a3a7](https://github.com/letta-ai/letta-node/commit/176a3a71f8c12afb5cf27a4e9913509a2f4a360c))
+* **internal:** update dependencies to address dependabot vulnerabilities ([2074604](https://github.com/letta-ai/letta-node/commit/20746042d8107a5fd2a5a59152c583593bade6a8))
+
+
+### Documentation
+
+* add OTID field description for idempotency context ([7c59ce1](https://github.com/letta-ai/letta-node/commit/7c59ce1f8c38d8172e44b91fafee619dd0403c0d))
+
+
+### Refactors
+
+* **sglang:** use input_ids instead of text in native adapter ([08679c1](https://github.com/letta-ai/letta-node/commit/08679c16d8ce0818fe8f72e77ae5dd366affc819))
+
 ## 1.7.12 (2026-03-09)
 
 Full Changelog: [v1.7.11...v1.7.12](https://github.com/letta-ai/letta-node/compare/v1.7.11...v1.7.12)

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@letta-ai/letta-client",
-  "version": "1.7.12",
+  "version": "1.8.1",
   "description": "The official TypeScript library for the Letta API",
   "author": "Letta <>",
   "types": "./index.d.ts",
@@ -25,6 +25,17 @@
     "fix": "./scripts/format"
   },
   "dependencies": {},
+  "overrides": {
+    "minimatch": "^9.0.5"
+  },
+  "pnpm": {
+    "overrides": {
+      "minimatch": "^9.0.5"
+    }
+  },
+  "resolutions": {
+    "minimatch": "^9.0.5"
+  },
   "exports": {
     ".": {
       "require": {

package/resources/agents/agents.d.mts

@@ -269,7 +269,7 @@ export interface AgentState {
     /**
      * The model settings used by the agent.
      */
-    model_settings?: OpenAIModelSettings | AnthropicModelSettings | GoogleAIModelSettings | GoogleVertexModelSettings | AzureModelSettings | XaiModelSettings | AgentState.ZaiModelSettings | GroqModelSettings | DeepseekModelSettings | TogetherModelSettings | BedrockModelSettings | AgentState.OpenRouterModelSettings | AgentState.ChatGptoAuthModelSettings | null;
+    model_settings?: OpenAIModelSettings | AgentState.SgLangModelSettings | AnthropicModelSettings | GoogleAIModelSettings | GoogleVertexModelSettings | AzureModelSettings | XaiModelSettings | AgentState.ZaiModelSettings | GroqModelSettings | DeepseekModelSettings | TogetherModelSettings | BedrockModelSettings | AgentState.OpenRouterModelSettings | AgentState.ChatGptoAuthModelSettings | null;
     /**
      * @deprecated Deprecated: Use `managed_group` field instead. The multi-agent group
      * that this agent manages.
@@ -525,7 +525,7 @@ export declare namespace AgentState {
         /**
          * Optional model settings used to override defaults for the summarizer model.
          */
-        model_settings?: AgentsAPI.OpenAIModelSettings | AgentsAPI.AnthropicModelSettings | AgentsAPI.GoogleAIModelSettings | AgentsAPI.GoogleVertexModelSettings | AgentsAPI.AzureModelSettings | AgentsAPI.XaiModelSettings | CompactionSettings.ZaiModelSettings | AgentsAPI.GroqModelSettings | AgentsAPI.DeepseekModelSettings | AgentsAPI.TogetherModelSettings | AgentsAPI.BedrockModelSettings | CompactionSettings.OpenRouterModelSettings | CompactionSettings.ChatGptoAuthModelSettings | null;
+        model_settings?: AgentsAPI.OpenAIModelSettings | CompactionSettings.SgLangModelSettings | AgentsAPI.AnthropicModelSettings | AgentsAPI.GoogleAIModelSettings | AgentsAPI.GoogleVertexModelSettings | AgentsAPI.AzureModelSettings | AgentsAPI.XaiModelSettings | CompactionSettings.ZaiModelSettings | AgentsAPI.GroqModelSettings | AgentsAPI.DeepseekModelSettings | AgentsAPI.TogetherModelSettings | AgentsAPI.BedrockModelSettings | CompactionSettings.OpenRouterModelSettings | CompactionSettings.ChatGptoAuthModelSettings | null;
         /**
          * The prompt to use for summarization. If None, uses mode-specific default.
          */
@@ -542,6 +542,56 @@ export declare namespace AgentState {
         sliding_window_percentage?: number;
     }
     namespace CompactionSettings {
+        /**
+         * SGLang model configuration (OpenAI-compatible runtime with SGLang-specific
+         * parsing).
+         */
+        interface SgLangModelSettings {
+            /**
+             * The maximum number of tokens the model can generate.
+             */
+            max_output_tokens?: number;
+            /**
+             * Whether to enable parallel tool calling.
+             */
+            parallel_tool_calls?: boolean;
+            /**
+             * The type of the provider.
+             */
+            provider_type?: 'sglang';
+            /**
+             * The reasoning configuration for the model.
+             */
+            reasoning?: SgLangModelSettings.Reasoning;
+            /**
+             * The response format for the model.
+             */
+            response_format?: AgentsAPI.TextResponseFormat | AgentsAPI.JsonSchemaResponseFormat | AgentsAPI.JsonObjectResponseFormat | null;
+            /**
+             * Enable strict mode for tool calling. When true, tool outputs are guaranteed to
+             * match JSON schemas.
+             */
+            strict?: boolean;
+            /**
+             * The temperature of the model.
+             */
+            temperature?: number;
+            /**
+             * SGLang tool call parser name (for example 'glm47', 'qwen25', or 'hermes').
+             */
+            tool_call_parser?: string | null;
+        }
+        namespace SgLangModelSettings {
+            /**
+             * The reasoning configuration for the model.
+             */
+            interface Reasoning {
+                /**
+                 * The reasoning effort to use when generating text reasoning models
+                 */
+                reasoning_effort?: 'none' | 'minimal' | 'low' | 'medium' | 'high' | 'xhigh';
+            }
+        }
         /**
          * Z.ai (ZhipuAI) model configuration (OpenAI-compatible).
          */
@@ -756,6 +806,56 @@ export declare namespace AgentState {
         termination_token?: string | null;
         turns_counter?: number | null;
     }
+    /**
+     * SGLang model configuration (OpenAI-compatible runtime with SGLang-specific
+     * parsing).
+     */
+    interface SgLangModelSettings {
+        /**
+         * The maximum number of tokens the model can generate.
+         */
+        max_output_tokens?: number;
+        /**
+         * Whether to enable parallel tool calling.
+         */
+        parallel_tool_calls?: boolean;
+        /**
+         * The type of the provider.
+         */
+        provider_type?: 'sglang';
+        /**
+         * The reasoning configuration for the model.
+         */
+        reasoning?: SgLangModelSettings.Reasoning;
+        /**
+         * The response format for the model.
+         */
+        response_format?: AgentsAPI.TextResponseFormat | AgentsAPI.JsonSchemaResponseFormat | AgentsAPI.JsonObjectResponseFormat | null;
+        /**
+         * Enable strict mode for tool calling. When true, tool outputs are guaranteed to
+         * match JSON schemas.
+         */
+        strict?: boolean;
+        /**
+         * The temperature of the model.
+         */
+        temperature?: number;
+        /**
+         * SGLang tool call parser name (for example 'glm47', 'qwen25', or 'hermes').
+         */
+        tool_call_parser?: string | null;
+    }
+    namespace SgLangModelSettings {
+        /**
+         * The reasoning configuration for the model.
+         */
+        interface Reasoning {
+            /**
+             * The reasoning effort to use when generating text reasoning models
+             */
+            reasoning_effort?: 'none' | 'minimal' | 'low' | 'medium' | 'high' | 'xhigh';
+        }
+    }
     /**
      * Z.ai (ZhipuAI) model configuration (OpenAI-compatible).
      */
@@ -1344,7 +1444,9 @@ export interface MessageCreate {
      */
     name?: string | null;
     /**
-     * The offline threading id associated with this message
+     * The offline threading id (OTID). Set by the client to deduplicate requests. Used
+     * for idempotency in background streaming mode — each message in a request must
+     * have a unique OTID. Retries of the same request should reuse the same OTIDs.
      */
     otid?: string | null;
     /**
@@ -1679,7 +1781,7 @@ export interface AgentCreateParams {
     /**
      * The model settings for the agent.
      */
-    model_settings?: OpenAIModelSettings | AnthropicModelSettings | GoogleAIModelSettings | GoogleVertexModelSettings | AzureModelSettings | XaiModelSettings | AgentCreateParams.ZaiModelSettings | GroqModelSettings | DeepseekModelSettings | TogetherModelSettings | BedrockModelSettings | AgentCreateParams.OpenRouterModelSettings | AgentCreateParams.ChatGptoAuthModelSettings | null;
+    model_settings?: OpenAIModelSettings | AgentCreateParams.SgLangModelSettings | AnthropicModelSettings | GoogleAIModelSettings | GoogleVertexModelSettings | AzureModelSettings | XaiModelSettings | AgentCreateParams.ZaiModelSettings | GroqModelSettings | DeepseekModelSettings | TogetherModelSettings | BedrockModelSettings | AgentCreateParams.OpenRouterModelSettings | AgentCreateParams.ChatGptoAuthModelSettings | null;
     /**
      * The name of the agent.
      */
@@ -1792,7 +1894,7 @@ export declare namespace AgentCreateParams {
         /**
          * Optional model settings used to override defaults for the summarizer model.
          */
-        model_settings?: AgentsAPI.OpenAIModelSettings | AgentsAPI.AnthropicModelSettings | AgentsAPI.GoogleAIModelSettings | AgentsAPI.GoogleVertexModelSettings | AgentsAPI.AzureModelSettings | AgentsAPI.XaiModelSettings | CompactionSettings.ZaiModelSettings | AgentsAPI.GroqModelSettings | AgentsAPI.DeepseekModelSettings | AgentsAPI.TogetherModelSettings | AgentsAPI.BedrockModelSettings | CompactionSettings.OpenRouterModelSettings | CompactionSettings.ChatGptoAuthModelSettings | null;
+        model_settings?: AgentsAPI.OpenAIModelSettings | CompactionSettings.SgLangModelSettings | AgentsAPI.AnthropicModelSettings | AgentsAPI.GoogleAIModelSettings | AgentsAPI.GoogleVertexModelSettings | AgentsAPI.AzureModelSettings | AgentsAPI.XaiModelSettings | CompactionSettings.ZaiModelSettings | AgentsAPI.GroqModelSettings | AgentsAPI.DeepseekModelSettings | AgentsAPI.TogetherModelSettings | AgentsAPI.BedrockModelSettings | CompactionSettings.OpenRouterModelSettings | CompactionSettings.ChatGptoAuthModelSettings | null;
         /**
          * The prompt to use for summarization. If None, uses mode-specific default.
          */
@@ -1809,6 +1911,56 @@ export declare namespace AgentCreateParams {
         sliding_window_percentage?: number;
     }
     namespace CompactionSettings {
+        /**
+         * SGLang model configuration (OpenAI-compatible runtime with SGLang-specific
+         * parsing).
+         */
+        interface SgLangModelSettings {
+            /**
+             * The maximum number of tokens the model can generate.
+             */
+            max_output_tokens?: number;
+            /**
+             * Whether to enable parallel tool calling.
+             */
+            parallel_tool_calls?: boolean;
+            /**
+             * The type of the provider.
+             */
+            provider_type?: 'sglang';
+            /**
+             * The reasoning configuration for the model.
+             */
+            reasoning?: SgLangModelSettings.Reasoning;
+            /**
+             * The response format for the model.
+             */
+            response_format?: AgentsAPI.TextResponseFormat | AgentsAPI.JsonSchemaResponseFormat | AgentsAPI.JsonObjectResponseFormat | null;
+            /**
+             * Enable strict mode for tool calling. When true, tool outputs are guaranteed to
+             * match JSON schemas.
+             */
+            strict?: boolean;
+            /**
+             * The temperature of the model.
+             */
+            temperature?: number;
+            /**
+             * SGLang tool call parser name (for example 'glm47', 'qwen25', or 'hermes').
+             */
+            tool_call_parser?: string | null;
+        }
+        namespace SgLangModelSettings {
+            /**
+             * The reasoning configuration for the model.
+             */
+            interface Reasoning {
+                /**
+                 * The reasoning effort to use when generating text reasoning models
+                 */
+                reasoning_effort?: 'none' | 'minimal' | 'low' | 'medium' | 'high' | 'xhigh';
+            }
+        }
         /**
          * Z.ai (ZhipuAI) model configuration (OpenAI-compatible).
          */
@@ -1915,6 +2067,56 @@ export declare namespace AgentCreateParams {
             }
         }
     }
+    /**
+     * SGLang model configuration (OpenAI-compatible runtime with SGLang-specific
+     * parsing).
+     */
+    interface SgLangModelSettings {
+        /**
+         * The maximum number of tokens the model can generate.
+         */
+        max_output_tokens?: number;
+        /**
+         * Whether to enable parallel tool calling.
+         */
+        parallel_tool_calls?: boolean;
+        /**
+         * The type of the provider.
+         */
+        provider_type?: 'sglang';
+        /**
+         * The reasoning configuration for the model.
+         */
+        reasoning?: SgLangModelSettings.Reasoning;
+        /**
+         * The response format for the model.
+         */
+        response_format?: AgentsAPI.TextResponseFormat | AgentsAPI.JsonSchemaResponseFormat | AgentsAPI.JsonObjectResponseFormat | null;
+        /**
+         * Enable strict mode for tool calling. When true, tool outputs are guaranteed to
+         * match JSON schemas.
+         */
+        strict?: boolean;
+        /**
+         * The temperature of the model.
+         */
+        temperature?: number;
+        /**
+         * SGLang tool call parser name (for example 'glm47', 'qwen25', or 'hermes').
+         */
+        tool_call_parser?: string | null;
+    }
+    namespace SgLangModelSettings {
+        /**
+         * The reasoning configuration for the model.
+         */
+        interface Reasoning {
+            /**
+             * The reasoning effort to use when generating text reasoning models
+             */
+            reasoning_effort?: 'none' | 'minimal' | 'low' | 'medium' | 'high' | 'xhigh';
+        }
+    }
     /**
      * Z.ai (ZhipuAI) model configuration (OpenAI-compatible).
      */
@@ -2137,7 +2339,7 @@ export interface AgentUpdateParams {
     /**
      * The model settings for the agent.
      */
-    model_settings?: OpenAIModelSettings | AnthropicModelSettings | GoogleAIModelSettings | GoogleVertexModelSettings | AzureModelSettings | XaiModelSettings | AgentUpdateParams.ZaiModelSettings | GroqModelSettings | DeepseekModelSettings | TogetherModelSettings | BedrockModelSettings | AgentUpdateParams.OpenRouterModelSettings | AgentUpdateParams.ChatGptoAuthModelSettings | null;
+    model_settings?: OpenAIModelSettings | AgentUpdateParams.SgLangModelSettings | AnthropicModelSettings | GoogleAIModelSettings | GoogleVertexModelSettings | AzureModelSettings | XaiModelSettings | AgentUpdateParams.ZaiModelSettings | GroqModelSettings | DeepseekModelSettings | TogetherModelSettings | BedrockModelSettings | AgentUpdateParams.OpenRouterModelSettings | AgentUpdateParams.ChatGptoAuthModelSettings | null;
     /**
      * The name of the agent.
      */
@@ -2233,7 +2435,7 @@ export declare namespace AgentUpdateParams {
         /**
          * Optional model settings used to override defaults for the summarizer model.
          */
-        model_settings?: AgentsAPI.OpenAIModelSettings | AgentsAPI.AnthropicModelSettings | AgentsAPI.GoogleAIModelSettings | AgentsAPI.GoogleVertexModelSettings | AgentsAPI.AzureModelSettings | AgentsAPI.XaiModelSettings | CompactionSettings.ZaiModelSettings | AgentsAPI.GroqModelSettings | AgentsAPI.DeepseekModelSettings | AgentsAPI.TogetherModelSettings | AgentsAPI.BedrockModelSettings | CompactionSettings.OpenRouterModelSettings | CompactionSettings.ChatGptoAuthModelSettings | null;
+        model_settings?: AgentsAPI.OpenAIModelSettings | CompactionSettings.SgLangModelSettings | AgentsAPI.AnthropicModelSettings | AgentsAPI.GoogleAIModelSettings | AgentsAPI.GoogleVertexModelSettings | AgentsAPI.AzureModelSettings | AgentsAPI.XaiModelSettings | CompactionSettings.ZaiModelSettings | AgentsAPI.GroqModelSettings | AgentsAPI.DeepseekModelSettings | AgentsAPI.TogetherModelSettings | AgentsAPI.BedrockModelSettings | CompactionSettings.OpenRouterModelSettings | CompactionSettings.ChatGptoAuthModelSettings | null;
         /**
          * The prompt to use for summarization. If None, uses mode-specific default.
          */
@@ -2250,6 +2452,56 @@ export declare namespace AgentUpdateParams {
         sliding_window_percentage?: number;
     }
     namespace CompactionSettings {
+        /**
+         * SGLang model configuration (OpenAI-compatible runtime with SGLang-specific
+         * parsing).
+         */
+        interface SgLangModelSettings {
+            /**
+             * The maximum number of tokens the model can generate.
+             */
+            max_output_tokens?: number;
+            /**
+             * Whether to enable parallel tool calling.
+             */
+            parallel_tool_calls?: boolean;
+            /**
+             * The type of the provider.
+             */
+            provider_type?: 'sglang';
+            /**
+             * The reasoning configuration for the model.
+             */
+            reasoning?: SgLangModelSettings.Reasoning;
+            /**
+             * The response format for the model.
+             */
+            response_format?: AgentsAPI.TextResponseFormat | AgentsAPI.JsonSchemaResponseFormat | AgentsAPI.JsonObjectResponseFormat | null;
+            /**
+             * Enable strict mode for tool calling. When true, tool outputs are guaranteed to
+             * match JSON schemas.
+             */
+            strict?: boolean;
+            /**
+             * The temperature of the model.
+             */
+            temperature?: number;
+            /**
+             * SGLang tool call parser name (for example 'glm47', 'qwen25', or 'hermes').
+             */
+            tool_call_parser?: string | null;
+        }
+        namespace SgLangModelSettings {
+            /**
+             * The reasoning configuration for the model.
+             */
+            interface Reasoning {
+                /**
+                 * The reasoning effort to use when generating text reasoning models
+                 */
+                reasoning_effort?: 'none' | 'minimal' | 'low' | 'medium' | 'high' | 'xhigh';
+            }
+        }
         /**
          * Z.ai (ZhipuAI) model configuration (OpenAI-compatible).
          */
@@ -2356,6 +2608,56 @@ export declare namespace AgentUpdateParams {
             }
         }
     }
+    /**
+     * SGLang model configuration (OpenAI-compatible runtime with SGLang-specific
+     * parsing).
+     */
+    interface SgLangModelSettings {
+        /**
+         * The maximum number of tokens the model can generate.
+         */
+        max_output_tokens?: number;
+        /**
+         * Whether to enable parallel tool calling.
+         */
+        parallel_tool_calls?: boolean;
+        /**
+         * The type of the provider.
+         */
+        provider_type?: 'sglang';
+        /**
+         * The reasoning configuration for the model.
+         */
+        reasoning?: SgLangModelSettings.Reasoning;
+        /**
+         * The response format for the model.
+         */
+        response_format?: AgentsAPI.TextResponseFormat | AgentsAPI.JsonSchemaResponseFormat | AgentsAPI.JsonObjectResponseFormat | null;
+        /**
+         * Enable strict mode for tool calling. When true, tool outputs are guaranteed to
+         * match JSON schemas.
+         */
+        strict?: boolean;
+        /**
+         * The temperature of the model.
+         */
+        temperature?: number;
+        /**
+         * SGLang tool call parser name (for example 'glm47', 'qwen25', or 'hermes').
+         */
+        tool_call_parser?: string | null;
+    }
+    namespace SgLangModelSettings {
+        /**
+         * The reasoning configuration for the model.
+         */
+        interface Reasoning {
+            /**
+             * The reasoning effort to use when generating text reasoning models
+             */
+            reasoning_effort?: 'none' | 'minimal' | 'low' | 'medium' | 'high' | 'xhigh';
+        }
+    }
     /**
      * Z.ai (ZhipuAI) model configuration (OpenAI-compatible).
      */
@@ -2472,6 +2774,10 @@ export interface AgentListParams extends ArrayPageParams {
      * Search agents by base template ID
      */
     base_template_id?: string | null;
+    /**
+     * Filter agents by the user who created them.
+     */
+    created_by_id?: string | null;
     /**
      * Search agents by identifier keys
      */