diff --git a/docs/docs/conduct/index.md b/docs/docs/conduct/index.md
index 404d4a6..febe0f8 100644
--- a/docs/docs/conduct/index.md
+++ b/docs/docs/conduct/index.md
@@ -4,3 +4,5 @@ How work is done in Comline's repositories.
 
 - [Branch rules](branch-rules.md): how changes reach a default branch, the
   required checks, review, and accepting a benchmark regression.
+- [Phases and promises](phases.md): early days, previews and the 1.0
+  announcement, and what is promised in each.
diff --git a/docs/docs/conduct/phases.md b/docs/docs/conduct/phases.md
new file mode 100644
index 0000000..077cf46
--- /dev/null
+++ b/docs/docs/conduct/phases.md
@@ -0,0 +1,57 @@
+# Phases and promises
+
+What Comline promises, and when. Decided 2026-10-27.
+
+## Phases
+
+- **Early days,** which is now.
+- **Previews,** smaller announcements before 1.0. Still early days.
+- **The 1.0 announcement,** which is what sets stability.
+
+## Always
+
+- Correctness, conformance, performance and safety are never lowered.
+- The [branch rules](branch-rules.md) and required checks apply.
+- A change to wire behaviour comes with conformance vectors, and every
+  runtime passes them.
+
+## Early days
+
+Until the 1.0 announcement.
+
+- Every version is `0.x`.
+- Anything may break:
+  - APIs and generated code;
+  - the C ABI, where a minor bump counts as breaking;
+  - the IR and config formats.
+- Wire formats (`packed@1`, `datagram@2`, `jsonrpc@1`, `comline@1`) may change
+  their bytes in place, but only in one cascade:
+  - the vectors change;
+  - every runtime (Rust, TypeScript, the simulator) passes them;
+  - all of it before anything merges.
+- No deprecation periods.
+- Everything is consumed by git revision, with pins such as `VECTORS_REV` and
+  the comline-core revision.
+- No registry publishing, no release workflow, no prebuilt binaries.
+- Proposals may still change.
+
+## Previews
+
+- Smaller announcements before 1.0.
+- They keep the early-days rules and say so: `0.x`, breaks allowed.
+
+## From the 1.0 announcement
+
+- Semver: a break needs a major version.
+- Anything users rely on is marked deprecated, with its replacement already
+  available, in at least one minor release before the major that removes it.
+  The changelog says so both times.
+- A wire format's bytes never change again. A change is a new `@N`.
+- The C ABI's structs are frozen within a major. Status codes and flags are
+  append-only.
+- Packages go to registries (npm, crates.io, …) under ComlineProject-owned
+  names.
+  - Published from CI with trusted publishing (OIDC), no long-lived tokens.
+  - Reproducible builds, with published checksums.
+  - Inter-repo development may keep using git revisions.
+- Background: [Ownership & packaging](../design/runtime-repo-structure.md#ownership-packaging).
diff --git a/docs/docs/design/runtime-repo-structure.md b/docs/docs/design/runtime-repo-structure.md
index de31db8..4341f4b 100644
--- a/docs/docs/design/runtime-repo-structure.md
+++ b/docs/docs/design/runtime-repo-structure.md
@@ -166,7 +166,8 @@ overhead with no payoff, so `comline-core` and `comline-codegen` are consumed by
 git rev — the way `generation` and `cli` already consume `core`. The cost is
 coordinated rev bumps across repos (and one `comline-core` in any given tree, so
 the CLI composes targets pinned to the same rev). Publishing to crates.io is a
-later call, not a prerequisite for anything below.
+later call, not a prerequisite for anything below; the
+[phases page](../conduct/phases.md) says when.
 
 ### Why the reversal
 
@@ -215,9 +216,10 @@ Steps 2–3 are the real prerequisites for a *second* target repo. `generation`'
 G3 in [Generation → G3](generation.md#g3-split-to-per-language-repos) is steps
 1 + 4 + 5 + 6 from the codegen side.
 
-**Publishing to crates.io** stays an open call for later. If the coordinated
-git-rev bumps across `comline-<lang>` repos get painful, that is the trigger to
-revisit — releasing `comline-core` and `comline-codegen` with semver is the fix,
+**Publishing to crates.io** stays an open call for later; registries come
+with the 1.0 rules on the [phases page](../conduct/phases.md). If the
+coordinated git-rev bumps across `comline-<lang>` repos get painful, that is the
+trigger to revisit — releasing `comline-core` and `comline-codegen` with semver is the fix,
 and it slots in without reordering anything here.
 
 ## Ownership & packaging
diff --git a/docs/docs/proposals/index.md b/docs/docs/proposals/index.md
index 7811c75..8084afd 100755
--- a/docs/docs/proposals/index.md
+++ b/docs/docs/proposals/index.md
@@ -74,7 +74,8 @@ reference. One proposal in three stages, read in order:
   handshake, framings and connection state machine behind a stable C ABI and
   WASM, built from one crate. ***accepted***, steps 1–3 built: the core
   changes, the C ABI at 0.1 ([reference](../reference/c-abi.md)), and its
-  vectors, fuzzing and benchmarks. Bindings and ABI 1.0 are next.
+  vectors, fuzzing and benchmarks. Bindings are next: TypeScript through wasm. The ABI stays `0.x` until the
+  [1.0 announcement](../conduct/phases.md).
 - [Stage 3 — generated codecs](shared-runtime-core/generated-codecs.md):
   native per-type codecs and protocol glue rendered from one plan in the
   codegen kernel. *under discussion.*
diff --git a/docs/docs/proposals/shared-runtime-core/core-interface.md b/docs/docs/proposals/shared-runtime-core/core-interface.md
index c1e8e88..2071fb6 100644
--- a/docs/docs/proposals/shared-runtime-core/core-interface.md
+++ b/docs/docs/proposals/shared-runtime-core/core-interface.md
@@ -160,8 +160,8 @@ else on this page is as proposed.
   The glue also reserves its regions on every call, not up front, and returns
   copies, so no view of wasm memory outlives a call.
 - **Deferred:**
-  - `comline_connection_size` and `comline_connection_init` come after 1.0, as a
-    minor addition;
+  - `comline_connection_size` and `comline_connection_init` are a later
+    addition;
   - `comline@2`'s negotiation comes before 1.0 if it changes the event struct.
 
 ### Found along the way
@@ -420,7 +420,7 @@ The FFI layer is for every other language, and Rust never pays for it.
     serde_json. See [Found along the way](#found-along-the-way).
 - A host with no allocator story of its own (embedded C, Zig) can get a
   later `comline_connection_size(cfg)` / `comline_connection_init(mem, …)`
-  pair. That is a minor ABI addition, **after 1.0**, not part of it.
+  pair. That is a later addition, not part of the ABI as built.
 
 **Panics never cross.**
 - Since Rust 1.81, a panic unwinding out of an `extern "C"` function aborts
@@ -468,6 +468,9 @@ The FFI layer is for every other language, and Rust never pays for it.
   first use.
 - **Before 1.0, a minor counts as a major.** A binding built against `0.n`
   accepts only `0.n`, so nothing pre-release is mistaken for compatible.
+- **The freeze comes with the project's 1.0 announcement,** not with a
+  milestone of the ABI's own ([phases](../../conduct/phases.md)). The ABI stays
+  `0.x` until then.
 - **Structs are frozen within a major.** New data comes through new
   functions or new flag bits. A `struct_size` field is not used, because
   every binding would have to honour it.
@@ -627,9 +630,11 @@ addition, not a replacement.
 
 ## Distribution
 
-**This is a sketch and an open question.** It needs its own pass before the
-release workflow and prebuilt binaries ([rollout](#rollout) step 5). Nothing
-in it is built.
+**This is a sketch and an open question.** Nothing in it is built.
+
+- Releases, registries and prebuilt binaries follow the post-1.0 rules on the
+  [phases page](../../conduct/phases.md).
+- Until then, bindings pin a runtime revision.
 
 **Targets.** Prebuilt binaries per triple, from `runtime`'s release
 workflow:
@@ -714,7 +719,7 @@ stays a separate idea.
 **Settled by the build:**
 
 - **Handle memory for no-allocator hosts:** `comline_connection_size` /
-  `comline_connection_init` come after 1.0, as a minor addition.
+  `comline_connection_init` are a later addition.
 - **`max_waiting` overflow:** an error to the caller of `encode_request`,
   `COMLINE_ERR_BUSY`. Queueing is the binding's job.
 - **Whether `jsonrpc@1`'s `serde_json` dependency fits the no-`std` FFI
@@ -725,8 +730,8 @@ stays a separate idea.
 **Still open:**
 
 - **Whether `comline@2`'s negotiation** (protocol numbers handed out at the
-  handshake) changes the event struct. If so, it should land before ABI 1.0
-  freezes.
+  handshake) changes the event struct. If so, it should land before 1.0,
+  when the ABI freezes.
 - **`pyo3` vs `cffi` for Python,** and whether a second binding surface is
   ever worth it. Decided by gates, but the default needs picking.
 - **Go without cgo:** wazero against pure Go, measured on the same
@@ -748,13 +753,12 @@ Steps 1–3 are done ([What's built](#whats-built)).
 3. **A C test binding** in `runtime` CI runs the conformance vectors through
    the C ABI and the wasm32 build: **done.** C and Node replay the same 65
    cases as Rust. This proves the boundary before any language depends on it.
-4. **The first two real bindings, chosen to stress the boundary:**
-   - **TypeScript**, through WASM and napi: stage 4's question;
-   - **Go**, through cgo: the most expensive call.
-   Both run the stage 1 gates against the TypeScript runtime and a pure-Go
-   baseline.
-5. **ABI 1.0** once both bindings pass, then the release workflow and
-   prebuilt binaries.
+4. **The first real binding: TypeScript, through WASM** (stage 4). Each
+   piece moves only where instruction counts show the core costs no more than
+   the TypeScript it replaces. napi and Go (cgo, the most expensive call)
+   come later, chosen by those numbers.
+5. **Bindings pin a runtime revision.** Releases and prebuilt binaries come
+   with the 1.0 rules ([phases](../../conduct/phases.md)).
 6. **Further languages** each bring a binding and a generator backend in
    their own repo.
 
diff --git a/docs/docs/proposals/shared-runtime-core/index.md b/docs/docs/proposals/shared-runtime-core/index.md
index 18bd420..6d792ac 100644
--- a/docs/docs/proposals/shared-runtime-core/index.md
+++ b/docs/docs/proposals/shared-runtime-core/index.md
@@ -112,7 +112,7 @@ Split the work by what can cross a language boundary cheaply.
 | Stage | Page | State |
 |---|---|---|
 | 1 | [Verification: differential fuzzing and benchmark gates](verification.md) | accepted, in progress |
-| 2 | [One protocol core: a C ABI and WASM from one crate](core-interface.md) | accepted; rollout steps 1–3 built, bindings and ABI 1.0 next; see the [C ABI reference](../../reference/c-abi.md) |
+| 2 | [One protocol core: a C ABI and WASM from one crate](core-interface.md) | accepted; rollout steps 1–3 built, bindings next: TypeScript through wasm; see the [C ABI reference](../../reference/c-abi.md) |
 | 3 | [Codecs and protocol glue generated from one plan](generated-codecs.md) | proposal |
 | 4 | Move TypeScript onto the core where the benchmarks allow | agreed |
 | 5 | Clean up the leftovers and the documents that contradict this | agreed |
@@ -129,6 +129,10 @@ rule CI enforces. Without it, stages 2–4 would be judged by opinion.
   vectors, fuzzing and gates.
 - The browser has no native option at all. The playground simulator already
   runs the core there.
+- The current plan is TypeScript through wasm only, each piece moving only
+  where instruction counts show the core is at least as fast. napi and Go
+  come later, chosen by those numbers. The C ABI stays `0.x` until the project
+  announces 1.0 ([phases](../../conduct/phases.md)).
 
 **Stage 5: the cleanup.**
 - **Remove or split `runtime/core/src/package_abi`.** It pulls `abi_stable`
diff --git a/docs/docs/reference/c-abi.md b/docs/docs/reference/c-abi.md
index aefa533..cc933b2 100644
--- a/docs/docs/reference/c-abi.md
+++ b/docs/docs/reference/c-abi.md
@@ -5,7 +5,8 @@ for the `comline@1` handshake, the connection framings (`datagram@2`,
 `jsonrpc@1`), envelopes, connection state and frame limits. Codecs,
 transports, async and dispatch stay in each host.
 
-- **ABI version 0.1.** It is not frozen; see [Version](#version).
+- **ABI version 0.1.** It is not frozen; it freezes when the project announces
+  1.0 ([phases](../conduct/phases.md)). See [Version](#version).
 - **Source:** [`runtime/ffi`](https://github.com/ComlineProject/runtime/tree/main/runtime/ffi)
   in the runtime repository.
 - **Header:** [`include/comline.h`](https://github.com/ComlineProject/runtime/blob/main/runtime/ffi/include/comline.h).
@@ -30,6 +31,8 @@ header disagree, the header is right.
 
 - **Before 1.0, a minor counts as a major.** A host built against `0.1` accepts
   only `0.1`. The JS glue does exactly this.
+- **The freeze comes with the project's 1.0 announcement,** not with an ABI
+  milestone of its own ([phases](../conduct/phases.md)).
 - The major changes when a signature, a struct layout or the meaning of a code
   changes. The minor changes when functions, codes or flag bits are added.
 - Structs are frozen within a major. New data comes through new functions or new
@@ -430,8 +433,9 @@ a 1% rise, like the core's benches. Both sides of each pair are gated.
 ## Not here yet
 
 - **`comline_connection_size` and `comline_connection_init`,** for a host that
-  supplies the handle's memory. After 1.0.
+  supplies the handle's memory. A later addition.
 - **`comline@2`'s negotiation** hands protocol numbers out at the handshake and
-  may change the event struct. If it does, it lands before 1.0 freezes the ABI.
+  may change the event struct. If it does, it lands before the 1.0
+  announcement freezes the ABI.
 - **Bindings.** No language binds the ABI yet; the JS glue is the only host
   code, and it is a thin wrapper, not a binding.
diff --git a/docs/mkdocs.yml b/docs/mkdocs.yml
index ffcc99f..7a846e1 100755
--- a/docs/mkdocs.yml
+++ b/docs/mkdocs.yml
@@ -147,6 +147,7 @@ nav:
   - Conduct:
     - 'conduct/index.md'
     - 'Branch rules': 'conduct/branch-rules.md'
+    - 'Phases and promises': 'conduct/phases.md'
 
 extra_css:
   - stylesheets/nav.css?v=21   # bump on every nav.css change — it is not cache-busted otherwise