diff --git a/Cargo.lock b/Cargo.lock
index 53a5bfe..9ee76d4 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -251,7 +251,7 @@ dependencies = [
 
 [[package]]
 name = "tcap"
-version = "0.1.1"
+version = "0.1.2"
 dependencies = [
  "anyhow",
  "clap",
diff --git a/Cargo.toml b/Cargo.toml
index 851fb3a..5804279 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "tcap"
-version = "0.1.1"
+version = "0.1.2"
 edition = "2021"
 description = "Capture the last command and its output from your terminal"
 license = "MIT"
diff --git a/README.md b/README.md
index 6c56e2e..ef41bc1 100644
--- a/README.md
+++ b/README.md
@@ -162,9 +162,8 @@ wrong block.
 **On kitty, output is the last *non-empty* output.** kitty counts the running
 `tcap` as the current command, so it has to be asked for the last output that
 wasn't empty. The consequence: if your last command printed nothing, kitty
-returns an *older* command's output while the header names the last one. The
-annotated header carries a `# note:` line saying so, and `--json` sets
-`"approximate": true`. Two `tcap` runs in a row are a **hard error** rather than
+returns an *older* command's output while the header names the last one. The annotated header carries a `# note:` line saying so and `--json` sets
+`"approximate": true`; a one-line reminder goes to stderr once per session. Two `tcap` runs in a row are a **hard error** rather than
 a label, since there the mismatch is certain. tmux has no such
 ambiguity, because the shell hook records real boundaries. kitty also
 needs remote control reachable — either `allow_remote_control yes`, or
diff --git a/npm/README.md b/npm/README.md
index 853364c..8d7f0a1 100644
--- a/npm/README.md
+++ b/npm/README.md
@@ -112,8 +112,9 @@ block.
 
 **On kitty, output is the last *non-empty* output**, because kitty counts the
 running `tcap` as the current command. A command that printed nothing therefore
-yields an older output under the current header; it is labelled with a `# note:`
-line and `"approximate": true` in `--json`. Two `tcap` runs in a row are a hard
+yields an older output under the current header; it is labelled with a `# note:` line and
+`"approximate": true` in `--json`, plus a one-line stderr reminder once per
+session. Two `tcap` runs in a row are a hard
 error instead, since the mismatch is certain there. tmux has no such ambiguity.
 
 **tmux wins whenever it's running.** Inside kitty running tmux, asking kitty for
diff --git a/npm/package.json b/npm/package.json
index 720365e..caa9e84 100644
--- a/npm/package.json
+++ b/npm/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@agora-build/tcap",
-  "version": "0.1.1",
+  "version": "0.1.2",
   "description": "Capture the last command and its output from your terminal, ready to pipe into an LLM",
   "bin": {
     "tcap": "bin/tcap"
diff --git a/src/main.rs b/src/main.rs
index 4d895fa..cacfc76 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -246,12 +246,16 @@ fn emit_hints(caps: &[Capture], backend: &dyn backend::Backend, mode: cli::Mode)
     //
     // The *certain* mismatch — a capture straight after another one — does not
     // reach here at all; capture_and_print refuses it outright.
-    if mode != cli::Mode::Command && caps.iter().any(|c| c.approximate) {
+    // Once per session, and one line. This is true of *every* kitty capture, so
+    // repeating it on each one is noise that buries the hints that are specific
+    // — and `tcap | llm` is the workflow it interrupted most.
+    if mode != cli::Mode::Command
+        && caps.iter().any(|c| c.approximate)
+        && state::first_time("approximate")
+    {
         eprintln!(
-            "tcap: {} returns the last non-empty output, which may not belong to the\n\
-             \x20 command named above — a silent command yields an earlier one's output,\n\
-             \x20 and a previous `tcap | …` leaves its pipeline's. tmux records exact\n\
-             \x20 boundaries; --quiet silences this.",
+            "tcap: {} cannot tie output to a command, so this may be an earlier one's \
+             — see `tcap doctor`. Shown once per session; --quiet silences it.",
             backend.name()
         );
         printed = true;
diff --git a/src/state.rs b/src/state.rs
index 4e72caa..4fde039 100644
--- a/src/state.rs
+++ b/src/state.rs
@@ -63,6 +63,20 @@ pub fn take_capture_marker() -> bool {
     fs::remove_file(marker_path()).is_ok()
 }
 
+/// True the first time it is asked for `key` in this session.
+///
+/// A caveat that is always true — kitty can never attribute output to a command
+/// — says nothing when repeated on every capture, and trains the reader to skip
+/// it. Said once, it still lands.
+pub fn first_time(key: &str) -> bool {
+    let path = state_dir().join(format!("{}.said-{key}", session_key()));
+    if fs::symlink_metadata(&path).is_ok() {
+        return false;
+    }
+    // If the note cannot be recorded, saying it again is the safer failure.
+    secure_write(&path, b"1").is_ok()
+}
+
 /// Why the state directory is unusable, if it is.
 ///
 /// `load` failing closed to an empty history also switches off the guard that