diff --git a/doc/extensions/strings.md b/doc/extensions/strings.md
new file mode 100644
index 00000000..3fafe8dc
--- /dev/null
+++ b/doc/extensions/strings.md
@@ -0,0 +1,60 @@
+<!--
+Copyright 2025 Google LLC
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    https://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+-->
+
+# Strings
+
+## string.format(list) -> string
+
+### Format
+
+`%[.precision]conversion`
+
+### Precision
+
+Optional. In the form of a period `.` followed by a required positive decimal digit sequence. The default precision is `6`. Not all conversions support precision.
+
+### Conversion
+
+|  Character | Precision | Description |
+| --- | --- | --- |
+| `s` | N | <table><tbody><tr><td><code>bool</code></td><td>The value is foramtted as <code>true</code> or <code>false</code>.</td></tr><tr><td><code>int</code></td><td>The value is formatted in base 10 with a preceding <code>-</code> if the value is negative. No insignificant <code>0</code>s must be included.</td></tr><tr><td><code>uint</code></td><td>The value is formatted in base 10. No insignificant <code>0</code>s must be included.</td></tr><tr><td><code>double</code></td><td>The value is formatted in base 10. No insignificant <code>0</code>s must be included. If there are no significant digits after the <code>.</code> then it must be excluded.</td></tr><tr><td><code>bytes</code></td><td>The value is formatted as if `string(value)` was performed and any invalid UTF-8 sequences are replaced with <code>\ufffd</code>. Multiple adjacent invalid UTF-8 sequences must be replaced with a single <code>\ufffd</code>.</td></tr><tr><td><code>string</code></td><td>The value is included as is.</td></tr><tr><td><code>duration</code></td><td>The value is formatted as decimal seconds as if the value was converted to <code>double</code> and then formatted as <code>%ds</code>.</td></tr><tr><td><code>timestamp</code></td><td>The value is formatted according to RFC 3339 and is always in UTC.</td></tr><tr><td><code>null_type</code></td><td>The value is formatted as <code>null</code>.</td></tr><tr><td><code>type</code></td><td>The value is formatted as a string.</td></tr><tr><td><code>list</code></td><td>The value is formatted as if each element was formatted as <code>"%s".format([element])</code>, joined together with <code>, </code> and enclosed with <code>[</code> and <code>]</code>.</td></tr><tr><td><code>map</code></td><td>The value is formatted as if each entry was formatted as <code>"%s: %s".format([key, value])</code>, sorted by the formatted keys in ascending order, joined together with <code>, </code>, and enclosed with <code>{</code> and <code>}</code>.</td></tr></tbody></table> |
+| `d` | N | <table><tbody><tr><td><code>int</code></td><td>The value is formatted in base 10 with a preceding <code>-</code> if the value is negative. No insignificant <code>0</code>s must be included.</td></tr><tr><td><code>uint</code></td><td>The value is formatted in base 10. No insignificant <code>0</code>s must be included.</td></tr><tr><td><code>double</code></td><td>The value is formatted in base 10. No insignificant <code>0</code>s must be included. If there are no significant digits after the <code>.</code> then it must be excluded.</td></tr></tbody></table> |
+| `f` | Y | `int` `uint` `double`: The value is converted to the style `[-]dddddd.dddddd` where there is at least one digit before the decimal and exactly `precision` digits after the decimal. If `precision` is 0, then the decimal is excluded. |
+| `e` | Y | `int` `uint` `double`: The value is converted to the style `[-]d.dddddde±dd` where there is one digit before the decimal and `precision` digits after the decimal followed by `e`, then the plus or minus, and then two digits. |
+| `x` `X` | N | Values are formatted in base 16. For `x` lowercase letters are used. For `X` uppercase letters are used.<table><tbody><tr><td><code>int</code> <code>uint</code></td><td>The value is formatted in base 16 with no insignificant digits. If the value was negative <code>-</code> is prepended.</td></tr><tr><td><code>string</code></td><td>The value is formatted as if `bytes(value)` was used to convert the <code>string</code> to <code>bytes</code> and then each byte is formatted in base 16 with exactly 2 digits.</td></tr><tr><td><code>bytes</code></td><td>The value is formatted as if each byte is formatted in base 16 with exactly 2 digits.</td></tr></tbody></table> |
+| `o` | N | `int` `uint`: The value is converted to base 8 with no insignificant digits. If the value was negative `-` is prepended. |
+| `b` | N | `int` `uint` `bool`: The value is converted to base 2 with no insignificant digits.  If the value was negative `-` is prepended. |
+
+> In all cases where `double` is accepted: if the value is NaN the result is `NaN`, if the value is infinity the result is `[-]Infinity`.
+
+### Examples
+
+```
+"%s".format(["foo"])  // foo
+"%s".format([b"foo"])  // foo
+"%d".format([1])      // 1
+"%d".format([1u])     // 1
+"%d".format([3.14])   // 3.14
+"%f".format([1])      // 1.000000
+"%f".format([1u])     // 1.000000
+"%f".format([3.14])   // 3.140000
+"%.1f".format([3.14])   // 3.1
+"%e".format([1])      // 1.000000e+00
+"%e".format([1u])     // 1.000000e+00
+"%e".format([3.14])   // 3.140000e+00
+"%.1e".format([3.14])   // 3.1e+00
+"%.1e".format([-3.14])   // -3.1e+00
+```
diff --git a/tests/simple/testdata/string_ext.textproto b/tests/simple/testdata/string_ext.textproto
index 67d1b606..eefa3bdb 100644
--- a/tests/simple/testdata/string_ext.textproto
+++ b/tests/simple/testdata/string_ext.textproto
@@ -524,7 +524,7 @@ section: {
     name: "scientific notation formatting clause"
     expr: '"%.6e".format([1052.032911275])'
     value: {
-      string_value: '1.052033×10⁰³',
+      string_value: '1.052033e+03',
     }
   }
   test: {
@@ -538,14 +538,49 @@ section: {
     name: "default precision for scientific notation"
     expr: '"%e".format([2.71828])'
     value: {
-      string_value: '2.718280×10⁰⁰',
+      string_value: '2.718280e+00',
     }
   }
   test: {
-    name: "unicode output for scientific notation"
-    expr: '"unescaped unicode: %e, escaped unicode: %e".format([2.71828, 2.71828])'
+    name: "NaN support for scientific notation"
+    expr: '"%e".format(["NaN"])'
     value: {
-      string_value: 'unescaped unicode: 2.718280×10⁰⁰, escaped unicode: 2.718280\u00d710\u2070\u2070',
+      string_value: 'NaN',
+    }
+  }
+  test: {
+    name: "positive infinity support for scientific notation"
+    expr: '"%e".format([double("Infinity")])'
+    value: {
+      string_value: 'Infinity',
+    }
+  }
+  test: {
+    name: "negative infinity support for scientific notation"
+    expr: '"%e".format([double("-Infinity")])'
+    value: {
+      string_value: '-Infinity',
+    }
+  }
+  test: {
+    name: "NaN support for decimal"
+    expr: '"%d".format(["NaN"])'
+    value: {
+      string_value: 'NaN',
+    }
+  }
+  test: {
+    name: "positive infinity support for decimal"
+    expr: '"%d".format([double("Infinity")])'
+    value: {
+      string_value: 'Infinity',
+    }
+  }
+  test: {
+    name: "negative infinity support for decimal"
+    expr: '"%d".format([double("-Infinity")])'
+    value: {
+      string_value: '-Infinity',
     }
   }
   test: {
@@ -557,16 +592,16 @@ section: {
   }
   test: {
     name: "positive infinity support for fixed-point"
-    expr: '"%f".format(["Infinity"])'
+    expr: '"%f".format([double("Infinity")])'
     value: {
-      string_value: '∞',
+      string_value: 'Infinity',
     }
   }
   test: {
     name: "negative infinity support for fixed-point"
-    expr: '"%f".format(["-Infinity"])'
+    expr: '"%f".format([double("-Infinity")])'
     value: {
-      string_value: '-∞',
+      string_value: '-Infinity',
     }
   }
   test: {
@@ -578,9 +613,9 @@ section: {
   }
   test: {
     name: "null support for string"
-    expr: '"null: %s".format([null])'
+    expr: '"%s".format([null])'
     value: {
-      string_value: 'null: null',
+      string_value: 'null',
     }
   }
   test: {
@@ -592,16 +627,16 @@ section: {
   }
   test: {
     name: "bytes support for string"
-    expr: '"some bytes: %s".format([b"xyz"])'
+    expr: '"%s".format([b"xyz"])'
     value: {
-      string_value: 'some bytes: xyz',
+      string_value: 'xyz',
     }
   }
   test: {
     name: "type() support for string"
-    expr: '"type is %s".format([type("test string")])'
+    expr: '"%s".format([type("test string")])'
     value: {
-      string_value: 'type is string',
+      string_value: 'string',
     }
   }
   test: {
@@ -620,135 +655,128 @@ section: {
   }
   test: {
     name: "list support for string"
-    expr: '"%s".format([["abc", 3.14, null, [9, 8, 7, 6], timestamp("2023-02-03T23:31:20Z")]])'
+    expr: '"%s".format([[abc, 3.14, null, [9, 8, 7, 6], 2023-02-03T23:31:20Z]])'
     value: {
-      string_value: '["abc", 3.14, null, [9, 8, 7, 6], timestamp("2023-02-03T23:31:20Z")]',
+      string_value: '[abc, 3.14, null, [9, 8, 7, 6], 2023-02-03T23:31:20Z]',
     }
   }
   test: {
     name: "map support for string"
     expr: '"%s".format([{"key1": b"xyz", "key5": null, "key2": duration("2h"), "key4": true, "key3": 2.71828}])'
     value: {
-      string_value: '{"key1":b"xyz", "key2":duration("7200s"), "key3":2.71828, "key4":true, "key5":null}',
+      string_value: '{key1: xyz, key2: 7200s, key3: 2.71828, key4: true, key5: null}',
     }
   }
   test: {
     name: "map support (all key types)"
-    expr: '"map with multiple key types: %s".format([{1: "value1", uint(2): "value2", true: double("NaN")}])'
+    expr: '"%s".format([{1: "value1", uint(2): "value2", true: double("NaN")}])'
     value: {
-      string_value: 'map with multiple key types: {1:"value1", 2:"value2", true:"NaN"}',
+      string_value: '{1: value1, 2: value2, true: NaN}',
     }
   }
   test: {
     name: "boolean support for %s"
-    expr: '"true bool: %s, false bool: %s".format([true, false])'
+    expr: '"%s, %s".format([true, false])'
     value: {
-      string_value: 'true bool: true, false bool: false',
+      string_value: 'true, false',
     }
   }
   test: {
       name: "dyntype support for string formatting clause"
-      expr: '"dynamic string: %s".format([dyn("a string")])'
+      expr: '"%s".format([dyn("a string")])'
       value: {
-          string_value: 'dynamic string: a string',
+          string_value: 'a string',
       }
   }
   test: {
     name: "dyntype support for numbers with string formatting clause"
-    expr: '"dynIntStr: %s dynDoubleStr: %s".format([dyn(32), dyn(56.8)])'
+    expr: '"%s, %s".format([dyn(32), dyn(56.8)])'
     value: {
-      string_value: 'dynIntStr: 32 dynDoubleStr: 56.8',
+      string_value: '32, 56.8',
     }
   }
   test: {
     name: "dyntype support for integer formatting clause"
-    expr: '"dynamic int: %d".format([dyn(128)])'
+    expr: '"%d".format([dyn(128)])'
     value: {
-      string_value: 'dynamic int: 128',
+      string_value: '128',
     }
   }
   test: {
     name: "dyntype support for integer formatting clause (unsigned)"
-    expr: '"dynamic unsigned int: %d".format([dyn(256u)])'
+    expr: '"%d".format([dyn(256u)])'
     value: {
-      string_value: 'dynamic unsigned int: 256',
+      string_value: '256',
     }
   }
   test: {
     name: "dyntype support for hex formatting clause"
-    expr: '"dynamic hex int: %x".format([dyn(22)])'
+    expr: '"%x".format([dyn(22)])'
     value: {
-      string_value: 'dynamic hex int: 16',
+      string_value: '16',
     }
   }
   test: {
     name: "dyntype support for hex formatting clause (uppercase)"
-    expr: '"dynamic hex int: %X (uppercase)".format([dyn(26)])'
+    expr: '"%X".format([dyn(26)])'
     value: {
-      string_value: 'dynamic hex int: 1A (uppercase)',
+      string_value: '1A',
     }
   }
   test: {
     name: "dyntype support for unsigned hex formatting clause"
-    expr: '"dynamic hex int: %x (unsigned)".format([dyn(500u)])'
+    expr: '"%x".format([dyn(500u)])'
     value: {
-      string_value: 'dynamic hex int: 1f4 (unsigned)',
+      string_value: '1f4',
     }
   }
   test: {
     name: "dyntype support for fixed-point formatting clause"
-    expr: '"dynamic double: %.3f".format([dyn(4.5)])'
+    expr: '"%.3f".format([dyn(4.5)])'
     value: {
-      string_value: 'dynamic double: 4.500',
+      string_value: '4.500',
     }
   }
   test: {
     name: "dyntype support for scientific notation"
-    expr: '"(dyntype) e: %e".format([dyn(2.71828)])'
+    expr: '"%e".format([dyn(2.71828)])'
     value: {
-      string_value: '(dyntype) e: 2.718280×10⁰⁰',
+      string_value: '2.718280e+00',
     }
   }
   test: {
     name: "dyntype NaN/infinity support for fixed-point"
-    expr: '"NaN: %f, infinity: %f".format([dyn("NaN"), dyn("Infinity")])'
+    expr: '"NaN: %f, infinity: %f".format([double("NaN"), double("Infinity"), double("-Infinity")])'
     value: {
-      string_value: 'NaN: NaN, infinity: ∞',
+      string_value: 'NaN, Infinity, -Infinity',
     }
   }
   test: {
     name: "dyntype support for timestamp"
-    expr: '"dyntype timestamp: %s".format([dyn(timestamp("2009-11-10T23:00:00Z"))])'
+    expr: '"%s".format([dyn(timestamp("2009-11-10T23:00:00Z"))])'
     value: {
-      string_value: 'dyntype timestamp: 2009-11-10T23:00:00Z',
+      string_value: '2009-11-10T23:00:00Z',
     }
   }
   test: {
     name: "dyntype support for duration"
-    expr: '"dyntype duration: %s".format([dyn(duration("8747s"))])'
+    expr: '"%s".format([dyn(duration("8747s"))])'
     value: {
-      string_value: 'dyntype duration: 8747s',
+      string_value: '8747s',
     }
   }
   test: {
     name: "dyntype support for lists"
-    expr: '"dyntype list: %s".format([dyn([6, 4.2, "a string"])])'
+    expr: '"%s".format([dyn([6, 4.2, "a string"])])'
     value: {
-      string_value: 'dyntype list: [6, 4.2, "a string"]',
+      string_value: '[6, 4.2, a string]',
     }
   }
   test: {
     name: "dyntype support for maps"
-    expr: '"dyntype map: %s".format([{"strKey":"x", 6:duration("422s"), true:42}])'
+    expr: '"%s".format([{"strKey":"x", 6:duration("422s"), true:42}])'
     value: {
-      string_value: 'dyntype map: {"strKey":"x", 6:duration("422s"), true:42}',
-    }
-  }
-  test: {
-    name: "message field support"
-    expr: '"message field msg.single_int32: %d, msg.single_double: %.1f".format([2, 1.0])'
-    value: {
-      string_value: 'message field msg.single_int32: 2, msg.single_double: 1.0',
+      string_value: '{strKey: x, 6: 422s, true: 42}',
     }
   }
   test: {
@@ -760,10 +788,10 @@ section: {
     }
     bindings: {
       key: "str_var"
-      value: { value: { string_value: "str is %s and some more" } }
+      value: { value: { string_value: "%s" } }
     }
     value: {
-      string_value: 'str is filler and some more',
+      string_value: 'filler',
     }
   }
   test: {
@@ -820,10 +848,10 @@ section: {
     }
     bindings: {
       key: "str_var"
-      value: { value: { string_value: "this is 5 in binary: %b" } }
+      value: { value: { string_value: "%b" } }
     }
     value: {
-      string_value: 'this is 5 in binary: 101',
+      string_value: '101',
     }
   }
   test: {
@@ -838,7 +866,7 @@ section: {
       value: { value: { string_value: "%.6e" } }
     }
     value: {
-      string_value: '1.052033×10⁰³',
+      string_value: '1.052033e+03',
     }
   }
   test: {