Never store requests (#3212)

When item ids are sent to Responses API it will load them from the database ignoring the provided values. This adds extra latency. Not having the mode to store requests also allows us to simplify the code. ## Breaking change The `disable_response_storage` configuration option is removed.
2025-09-05 10:41:47 -07:00
parent ba631e7928
commit 5775174ec2
17 changed files with 12 additions and 118 deletions
--- a/codex-rs/core/src/client.rs
+++ b/codex-rs/core/src/client.rs
@@ -157,14 +157,6 @@ impl ModelClient {

        let auth_manager = self.auth_manager.clone();

-        let auth_mode = auth_manager
-            .as_ref()
-            .and_then(|m| m.auth())
-            .as_ref()
-            .map(|a| a.mode);
-
-        let store = prompt.store && auth_mode != Some(AuthMode::ChatGPT);
-
        let full_instructions = prompt.get_full_instructions(&self.config.model_family);
        let tools_json = create_tools_json_for_responses_api(&prompt.tools)?;
        let reasoning = create_reasoning_param_for_request(
@@ -173,9 +165,7 @@ impl ModelClient {
            self.summary,
        );

-        // Request encrypted COT if we are not storing responses,
-        // otherwise reasoning items will be referenced by ID
-        let include: Vec<String> = if !store && reasoning.is_some() {
+        let include: Vec<String> = if reasoning.is_some() {
            vec!["reasoning.encrypted_content".to_string()]
        } else {
            vec![]
@@ -204,7 +194,7 @@ impl ModelClient {
            tool_choice: "auto",
            parallel_tool_calls: false,
            reasoning,
-            store,
+            store: false,
            stream: true,
            include,
            prompt_cache_key: Some(self.session_id.to_string()),
--- a/codex-rs/core/src/client_common.rs
+++ b/codex-rs/core/src/client_common.rs
@@ -25,9 +25,6 @@ pub struct Prompt {
    /// Conversation context input items.
    pub input: Vec<ResponseItem>,

-    /// Whether to store response on server side (disable_response_storage = !store).
-    pub store: bool,
-
    /// Tools available to the model, including additional tools sourced from
    /// external MCP servers.
    pub(crate) tools: Vec<OpenAiTool>,
@@ -128,7 +125,6 @@ pub(crate) struct ResponsesApiRequest<'a> {
    pub(crate) tool_choice: &'static str,
    pub(crate) parallel_tool_calls: bool,
    pub(crate) reasoning: Option<Reasoning>,
-    /// true when using the Responses API.
    pub(crate) store: bool,
    pub(crate) stream: bool,
    pub(crate) include: Vec<String>,
@@ -199,7 +195,7 @@ mod tests {
            tool_choice: "auto",
            parallel_tool_calls: false,
            reasoning: None,
-            store: true,
+            store: false,
            stream: true,
            include: vec![],
            prompt_cache_key: None,
@@ -229,7 +225,7 @@ mod tests {
            tool_choice: "auto",
            parallel_tool_calls: false,
            reasoning: None,
-            store: true,
+            store: false,
            stream: true,
            include: vec![],
            prompt_cache_key: None,
--- a/codex-rs/core/src/codex.rs
+++ b/codex-rs/core/src/codex.rs
@@ -184,7 +184,6 @@ impl Codex {
            base_instructions: config.base_instructions.clone(),
            approval_policy: config.approval_policy,
            sandbox_policy: config.sandbox_policy.clone(),
-            disable_response_storage: config.disable_response_storage,
            notify: config.notify.clone(),
            cwd: config.cwd.clone(),
        };
@@ -301,7 +300,6 @@ pub(crate) struct TurnContext {
    pub(crate) approval_policy: AskForApproval,
    pub(crate) sandbox_policy: SandboxPolicy,
    pub(crate) shell_environment_policy: ShellEnvironmentPolicy,
-    pub(crate) disable_response_storage: bool,
    pub(crate) tools_config: ToolsConfig,
 }

@@ -334,8 +332,6 @@ struct ConfigureSession {
    approval_policy: AskForApproval,
    /// How to sandbox commands executed in the system
    sandbox_policy: SandboxPolicy,
-    /// Disable server-side response storage (send full context each request)
-    disable_response_storage: bool,

    /// Optional external notifier command tokens. Present only when the
    /// client wants the agent to spawn a program after each completed
@@ -370,7 +366,6 @@ impl Session {
            base_instructions,
            approval_policy,
            sandbox_policy,
-            disable_response_storage,
            notify,
            cwd,
        } = configure_session;
@@ -462,7 +457,6 @@ impl Session {
            sandbox_policy,
            shell_environment_policy: config.shell_environment_policy.clone(),
            cwd,
-            disable_response_storage,
        };
        let sess = Arc::new(Session {
            session_id,
@@ -1117,7 +1111,6 @@ async fn submission_loop(
                    sandbox_policy: new_sandbox_policy.clone(),
                    shell_environment_policy: prev.shell_environment_policy.clone(),
                    cwd: new_cwd.clone(),
-                    disable_response_storage: prev.disable_response_storage,
                };

                // Install the new persistent context for subsequent tasks/turns.
@@ -1199,7 +1192,6 @@ async fn submission_loop(
                        sandbox_policy,
                        shell_environment_policy: turn_context.shell_environment_policy.clone(),
                        cwd,
-                        disable_response_storage: turn_context.disable_response_storage,
                    };
                    // TODO: record the new environment context in the conversation history
                    // no current task, spawn a new one with the per‑turn context
@@ -1604,7 +1596,6 @@ async fn run_turn(

    let prompt = Prompt {
        input,
-        store: !turn_context.disable_response_storage,
        tools,
        base_instructions_override: turn_context.base_instructions.clone(),
    };
@@ -1858,7 +1849,6 @@ async fn run_compact_task(

    let prompt = Prompt {
        input: turn_input,
-        store: !turn_context.disable_response_storage,
        tools: Vec::new(),
        base_instructions_override: Some(compact_instructions.clone()),
    };
--- a/codex-rs/core/src/config.rs
+++ b/codex-rs/core/src/config.rs
@@ -78,11 +78,6 @@ pub struct Config {
    /// Defaults to `false`.
    pub show_raw_agent_reasoning: bool,

-    /// Disable server-side response storage (sends the full conversation
-    /// context with every request). Currently necessary for OpenAI customers
-    /// who have opted into Zero Data Retention (ZDR).
-    pub disable_response_storage: bool,
-
    /// User-provided instructions from AGENTS.md.
    pub user_instructions: Option<String>,

@@ -417,11 +412,6 @@ pub struct ConfigToml {
    /// Sandbox configuration to apply if `sandbox` is `WorkspaceWrite`.
    pub sandbox_workspace_write: Option<SandboxWorkspaceWrite>,

-    /// Disable server-side response storage (sends the full conversation
-    /// context with every request). Currently necessary for OpenAI customers
-    /// who have opted into Zero Data Retention (ZDR).
-    pub disable_response_storage: Option<bool>,
-
    /// Optional external command to spawn for end-user notifications.
    #[serde(default)]
    pub notify: Option<Vec<String>>,
@@ -640,7 +630,6 @@ pub struct ConfigOverrides {
    pub include_plan_tool: Option<bool>,
    pub include_apply_patch_tool: Option<bool>,
    pub include_view_image_tool: Option<bool>,
-    pub disable_response_storage: Option<bool>,
    pub show_raw_agent_reasoning: Option<bool>,
    pub tools_web_search_request: Option<bool>,
 }
@@ -668,7 +657,6 @@ impl Config {
            include_plan_tool,
            include_apply_patch_tool,
            include_view_image_tool,
-            disable_response_storage,
            show_raw_agent_reasoning,
            tools_web_search_request: override_tools_web_search_request,
        } = overrides;
@@ -802,11 +790,6 @@ impl Config {
                .unwrap_or_else(AskForApproval::default),
            sandbox_policy,
            shell_environment_policy,
-            disable_response_storage: config_profile
-                .disable_response_storage
-                .or(cfg.disable_response_storage)
-                .or(disable_response_storage)
-                .unwrap_or(false),
            notify: cfg.notify,
            user_instructions,
            base_instructions,
@@ -1071,7 +1054,6 @@ exclude_slash_tmp = true
        let toml = r#"
 model = "o3"
 approval_policy = "untrusted"
-disable_response_storage = false

 # Can be used to determine which profile to use if not specified by
 # `ConfigOverrides`.
@@ -1101,7 +1083,6 @@ model_provider = "openai-chat-completions"
 model = "o3"
 model_provider = "openai"
 approval_policy = "on-failure"
-disable_response_storage = true

 [profiles.gpt5]
 model = "gpt-5"
@@ -1199,7 +1180,6 @@ model_verbosity = "high"
                approval_policy: AskForApproval::Never,
                sandbox_policy: SandboxPolicy::new_read_only_policy(),
                shell_environment_policy: ShellEnvironmentPolicy::default(),
-                disable_response_storage: false,
                user_instructions: None,
                notify: None,
                cwd: fixture.cwd(),
@@ -1257,7 +1237,6 @@ model_verbosity = "high"
            approval_policy: AskForApproval::UnlessTrusted,
            sandbox_policy: SandboxPolicy::new_read_only_policy(),
            shell_environment_policy: ShellEnvironmentPolicy::default(),
-            disable_response_storage: false,
            user_instructions: None,
            notify: None,
            cwd: fixture.cwd(),
@@ -1330,7 +1309,6 @@ model_verbosity = "high"
            approval_policy: AskForApproval::OnFailure,
            sandbox_policy: SandboxPolicy::new_read_only_policy(),
            shell_environment_policy: ShellEnvironmentPolicy::default(),
-            disable_response_storage: true,
            user_instructions: None,
            notify: None,
            cwd: fixture.cwd(),
@@ -1389,7 +1367,6 @@ model_verbosity = "high"
            approval_policy: AskForApproval::OnFailure,
            sandbox_policy: SandboxPolicy::new_read_only_policy(),
            shell_environment_policy: ShellEnvironmentPolicy::default(),
-            disable_response_storage: false,
            user_instructions: None,
            notify: None,
            cwd: fixture.cwd(),
--- a/codex-rs/core/src/config_profile.rs
+++ b/codex-rs/core/src/config_profile.rs
@@ -15,7 +15,6 @@ pub struct ConfigProfile {
    /// [`ModelProviderInfo`] to use.
    pub model_provider: Option<String>,
    pub approval_policy: Option<AskForApproval>,
-    pub disable_response_storage: Option<bool>,
    pub model_reasoning_effort: Option<ReasoningEffort>,
    pub model_reasoning_summary: Option<ReasoningSummary>,
    pub model_verbosity: Option<Verbosity>,
@@ -29,7 +28,6 @@ impl From<ConfigProfile> for codex_protocol::mcp_protocol::Profile {
            model: config_profile.model,
            model_provider: config_profile.model_provider,
            approval_policy: config_profile.approval_policy,
-            disable_response_storage: config_profile.disable_response_storage,
            model_reasoning_effort: config_profile.model_reasoning_effort,
            model_reasoning_summary: config_profile.model_reasoning_summary,
            model_verbosity: config_profile.model_verbosity,
--- a/codex-rs/core/tests/suite/client.rs
+++ b/codex-rs/core/tests/suite/client.rs
@@ -224,19 +224,16 @@ async fn resume_includes_initial_messages_and_sends_prior_items() {
    let expected_input = serde_json::json!([
        {
            "type": "message",
-            "id": null,
            "role": "user",
            "content": [{ "type": "input_text", "text": "resumed user message" }]
        },
        {
            "type": "message",
-            "id": null,
            "role": "assistant",
            "content": [{ "type": "output_text", "text": "resumed assistant message" }]
        },
        {
            "type": "message",
-            "id": null,
            "role": "user",
            "content": [{ "type": "input_text", "text": "hello" }]
        }
@@ -496,7 +493,6 @@ async fn chatgpt_auth_sends_correct_request() {
        "Bearer Access Token"
    );
    assert_eq!(request_chatgpt_account_id.to_str().unwrap(), "account_id");
-    assert!(!request_body["store"].as_bool().unwrap());
    assert!(request_body["stream"].as_bool().unwrap());
    assert_eq!(
        request_body["include"][0].as_str().unwrap(),
@@ -578,14 +574,6 @@ async fn prefers_chatgpt_token_when_config_prefers_chatgpt() {
        .unwrap();

    wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
-
-    // verify request body flags
-    let request = &server.received_requests().await.unwrap()[0];
-    let request_body = request.body_json::<serde_json::Value>().unwrap();
-    assert!(
-        !request_body["store"].as_bool().unwrap(),
-        "store should be false for ChatGPT auth"
-    );
 }

 #[tokio::test(flavor = "multi_thread", worker_threads = 2)]
@@ -662,14 +650,6 @@ async fn prefers_apikey_when_config_prefers_apikey_even_with_chatgpt_tokens() {
        .unwrap();

    wait_for_event(&codex, |ev| matches!(ev, EventMsg::TaskComplete(_))).await;
-
-    // verify request body flags
-    let request = &server.received_requests().await.unwrap()[0];
-    let request_body = request.body_json::<serde_json::Value>().unwrap();
-    assert!(
-        request_body["store"].as_bool().unwrap(),
-        "store should be true for API key auth"
-    );
 }

 #[tokio::test(flavor = "multi_thread", worker_threads = 2)]
@@ -990,31 +970,26 @@ async fn history_dedupes_streamed_and_final_messages_across_turns() {
    let r3_tail_expected = serde_json::json!([
        {
            "type": "message",
-            "id": null,
            "role": "user",
            "content": [{"type":"input_text","text":"U1"}]
        },
        {
            "type": "message",
-            "id": null,
            "role": "assistant",
            "content": [{"type":"output_text","text":"Hey there!\n"}]
        },
        {
            "type": "message",
-            "id": null,
            "role": "user",
            "content": [{"type":"input_text","text":"U2"}]
        },
        {
            "type": "message",
-            "id": null,
            "role": "assistant",
            "content": [{"type":"output_text","text":"Hey there!\n"}]
        },
        {
            "type": "message",
-            "id": null,
            "role": "user",
            "content": [{"type":"input_text","text":"U3"}]
        }
--- a/codex-rs/core/tests/suite/prompt_caching.rs
+++ b/codex-rs/core/tests/suite/prompt_caching.rs
@@ -289,20 +289,17 @@ async fn prefixes_context_and_instructions_once_and_consistently_across_requests

    let expected_env_msg = serde_json::json!({
        "type": "message",
-        "id": serde_json::Value::Null,
        "role": "user",
        "content": [ { "type": "input_text", "text": expected_env_text } ]
    });
    let expected_ui_msg = serde_json::json!({
        "type": "message",
-        "id": serde_json::Value::Null,
        "role": "user",
        "content": [ { "type": "input_text", "text": expected_ui_text } ]
    });

    let expected_user_message_1 = serde_json::json!({
        "type": "message",
-        "id": serde_json::Value::Null,
        "role": "user",
        "content": [ { "type": "input_text", "text": "hello 1" } ]
    });
@@ -314,7 +311,6 @@ async fn prefixes_context_and_instructions_once_and_consistently_across_requests

    let expected_user_message_2 = serde_json::json!({
        "type": "message",
-        "id": serde_json::Value::Null,
        "role": "user",
        "content": [ { "type": "input_text", "text": "hello 2" } ]
    });
@@ -424,7 +420,6 @@ async fn overrides_turn_context_but_keeps_cached_prefix_and_key_constant() {
    // as the prefix of the second request, ensuring cache hit potential.
    let expected_user_message_2 = serde_json::json!({
        "type": "message",
-        "id": serde_json::Value::Null,
        "role": "user",
        "content": [ { "type": "input_text", "text": "hello 2" } ]
    });
@@ -438,7 +433,6 @@ async fn overrides_turn_context_but_keeps_cached_prefix_and_key_constant() {
 </environment_context>"#;
    let expected_env_msg_2 = serde_json::json!({
        "type": "message",
-        "id": serde_json::Value::Null,
        "role": "user",
        "content": [ { "type": "input_text", "text": expected_env_text_2 } ]
    });
@@ -543,7 +537,6 @@ async fn per_turn_overrides_keep_cached_prefix_and_key_constant() {
    // as the prefix of the second request.
    let expected_user_message_2 = serde_json::json!({
        "type": "message",
-        "id": serde_json::Value::Null,
        "role": "user",
        "content": [ { "type": "input_text", "text": "hello 2" } ]
    });
--- a/codex-rs/exec/src/lib.rs
+++ b/codex-rs/exec/src/lib.rs
@@ -149,7 +149,6 @@ pub async fn run_main(cli: Cli, codex_linux_sandbox_exe: Option<PathBuf>) -> any
        include_plan_tool: None,
        include_apply_patch_tool: None,
        include_view_image_tool: None,
-        disable_response_storage: oss.then_some(true),
        show_raw_agent_reasoning: oss.then_some(true),
        tools_web_search_request: None,
    };
--- a/codex-rs/mcp-server/src/codex_message_processor.rs
+++ b/codex-rs/mcp-server/src/codex_message_processor.rs
@@ -994,7 +994,6 @@ fn derive_config_from_params(
        include_plan_tool,
        include_apply_patch_tool,
        include_view_image_tool: None,
-        disable_response_storage: None,
        show_raw_agent_reasoning: None,
        tools_web_search_request: None,
    };
--- a/codex-rs/mcp-server/src/codex_tool_config.rs
+++ b/codex-rs/mcp-server/src/codex_tool_config.rs
@@ -162,7 +162,6 @@ impl CodexToolCallParam {
            include_plan_tool,
            include_apply_patch_tool: None,
            include_view_image_tool: None,
-            disable_response_storage: None,
            show_raw_agent_reasoning: None,
            tools_web_search_request: None,
        };
--- a/codex-rs/mcp-server/tests/suite/config.rs
+++ b/codex-rs/mcp-server/tests/suite/config.rs
@@ -51,7 +51,6 @@ model_reasoning_effort = "high"
 model_reasoning_summary = "detailed"
 model_verbosity = "medium"
 model_provider = "openai"
-disable_response_storage = false
 chatgpt_base_url = "https://api.chatgpt.com"
 "#,
    )
@@ -111,7 +110,6 @@ async fn get_config_toml_parses_all_fields() {
                    model_reasoning_summary: Some(ReasoningSummary::Detailed),
                    model_verbosity: Some(Verbosity::Medium),
                    model_provider: Some("openai".into()),
-                    disable_response_storage: Some(false),
                    chatgpt_base_url: Some("https://api.chatgpt.com".into()),
                },
            )]),
--- a/codex-rs/protocol/src/mcp_protocol.rs
+++ b/codex-rs/protocol/src/mcp_protocol.rs
@@ -371,7 +371,6 @@ pub struct Profile {
    /// [`ModelProviderInfo`] to use.
    pub model_provider: Option<String>,
    pub approval_policy: Option<AskForApproval>,
-    pub disable_response_storage: Option<bool>,
    pub model_reasoning_effort: Option<ReasoningEffort>,
    pub model_reasoning_summary: Option<ReasoningSummary>,
    pub model_verbosity: Option<Verbosity>,
--- a/codex-rs/protocol/src/models.rs
+++ b/codex-rs/protocol/src/models.rs
@@ -42,11 +42,13 @@ pub enum ContentItem {
 #[serde(tag = "type", rename_all = "snake_case")]
 pub enum ResponseItem {
    Message {
+        #[serde(skip_serializing)]
        id: Option<String>,
        role: String,
        content: Vec<ContentItem>,
    },
    Reasoning {
+        #[serde(default)]
        id: String,
        summary: Vec<ReasoningItemReasoningSummary>,
        #[serde(default, skip_serializing_if = "should_serialize_reasoning_content")]
@@ -55,6 +57,7 @@ pub enum ResponseItem {
    },
    LocalShellCall {
        /// Set when using the chat completions API.
+        #[serde(skip_serializing)]
        id: Option<String>,
        /// Set when using the Responses API.
        call_id: Option<String>,
@@ -62,6 +65,7 @@ pub enum ResponseItem {
        action: LocalShellAction,
    },
    FunctionCall {
+        #[serde(skip_serializing)]
        id: Option<String>,
        name: String,
        // The Responses API returns the function call arguments as a *string* that contains
@@ -82,7 +86,7 @@ pub enum ResponseItem {
        output: FunctionCallOutputPayload,
    },
    CustomToolCall {
-        #[serde(default, skip_serializing_if = "Option::is_none")]
+        #[serde(skip_serializing)]
        id: Option<String>,
        #[serde(default, skip_serializing_if = "Option::is_none")]
        status: Option<String>,
@@ -104,7 +108,7 @@ pub enum ResponseItem {
    //   "action": {"type":"search","query":"weather: San Francisco, CA"}
    // }
    WebSearchCall {
-        #[serde(default, skip_serializing_if = "Option::is_none")]
+        #[serde(skip_serializing)]
        id: Option<String>,
        #[serde(default, skip_serializing_if = "Option::is_none")]
        status: Option<String>,
--- a/codex-rs/tui/src/lib.rs
+++ b/codex-rs/tui/src/lib.rs
@@ -130,7 +130,6 @@ pub async fn run_main(
        include_plan_tool: Some(true),
        include_apply_patch_tool: None,
        include_view_image_tool: None,
-        disable_response_storage: cli.oss.then_some(true),
        show_raw_agent_reasoning: cli.oss.then_some(true),
        tools_web_search_request: cli.web_search.then_some(true),
    };
--- a/docs/config.md
+++ b/docs/config.md
@@ -182,7 +182,6 @@ Here is an example of a `config.toml` that defines multiple profiles:
 ```toml
 model = "o3"
 approval_policy = "untrusted"
-disable_response_storage = false

 # Setting `profile` is equivalent to specifying `--profile o3` on the command
 # line, though the `--profile` flag can still be used to override this value.
@@ -209,7 +208,6 @@ model_provider = "openai-chat-completions"
 model = "o3"
 model_provider = "openai"
 approval_policy = "on-failure"
-disable_response_storage = true
 ```

 Users can specify config values at multiple levels. Order of precedence is as follows:
@@ -362,14 +360,6 @@ args = ["-y", "mcp-server"]
 env = { "API_KEY" = "value" }
 ```

-## disable_response_storage
-
-Currently, customers whose accounts are set to use Zero Data Retention (ZDR) must set `disable_response_storage` to `true` so that Codex uses an alternative to the Responses API that works with ZDR:
-
-```toml
-disable_response_storage = true
-```
-
 ## shell_environment_policy

 Codex spawns subprocesses (e.g. when executing a `local_shell` tool-call suggested by the assistant). By default it now passes **your full environment** to those subprocesses. You can tune this behavior via the **`shell_environment_policy`** block in `config.toml`:
--- a/docs/zdr.md
+++ b/docs/zdr.md
@@ -1,15 +1,3 @@
 ## Zero data retention (ZDR) usage

-Codex CLI **does** support OpenAI organizations with [Zero Data Retention (ZDR)](https://platform.openai.com/docs/guides/your-data#zero-data-retention) enabled. If your OpenAI organization has Zero Data Retention enabled and you still encounter errors such as:
-
-```
-OpenAI rejected the request. Error details: Status: 400, Code: unsupported_parameter, Type: invalid_request_error, Message: 400 Previous response cannot be used for this organization due to Zero Data Retention.
-```
-
-Ensure you are running `codex` with `--config disable_response_storage=true` or add this line to `~/.codex/config.toml` to avoid specifying the command line option each time:
-
-```toml
-disable_response_storage = true
-```
-
-See [the configuration documentation on `disable_response_storage`](./config.md#disable_response_storage) for details. 
+Codex CLI natively supports OpenAI organizations with [Zero Data Retention (ZDR)](https://platform.openai.com/docs/guides/your-data#zero-data-retention) enabled.