diff --git a/packages/api/README.md b/packages/api/README.md
index 112818d..b075c03 100644
--- a/packages/api/README.md
+++ b/packages/api/README.md
@@ -81,26 +81,17 @@ Jetstream's words, delivery is at-least-once, so replaying the same create or th
 twice leaves the index unchanged — because the whole index can be thrown away and rebuilt from
 the network at any time.
 
-### Deviation from the draft spec: Jetstream's live API has moved
-
-The draft spec (`api-spec.md`) assumed Jetstream's legacy `/subscribe` endpoint
-(`wantedCollections=...&cursor=<time_us>`). Reading the current docs
-(bsky.network/docs/jetstream, 2026-11-06) instead of relying on memory, as the task asked, found
-a v2 API: `wss://<host>/xrpc/network.bsky.jetstream.subscribeEvents`, with a repeated
-`collections` param (confirmed live, documented as supporting an
-`app.bsky.feed.*`-style NSID-prefix wildcard, so `run.nibling.*` would also work), `kinds`, and a
-`cursor` that takes either a `seq` or a `time_us`, told apart by magnitude. The index stores a
-`time_us`, since that also gives the cursor's age; a live probe on 2026-11-06 with a `time_us`
-two minutes old replayed from two minutes back. `src/index/jetstream.ts` targets the verified v2 shape; `wrangler.jsonc`'s
-`JETSTREAM_URL` points at `wss://jetstream.us-east.bsky.network` accordingly. A live, read-only
-connection to that URL with this worker's exact query string opened cleanly and returned zero
-`run.nibling.*` commits in an 8-second window, which is expected: nibling.run has no real users
-yet.
-
-`com.atproto.sync.listReposByCollection` itself matched the spec and current docs exactly
-(`collection`, `limit`, `cursor` in; `{ repos: [{did}], cursor }` out); a live, read-only request
-against `relay1.us-east.bsky.network` for `run.nibling.project` returned `{"repos":[]}`, as
-expected for a collection nobody writes to yet.
+### Jetstream v2
+
+Jetstream's live API is `wss://<host>/xrpc/network.bsky.jetstream.subscribeEvents`, with a
+repeated `collections` param (an `app.bsky.feed.*`-style NSID-prefix wildcard works, so
+`run.nibling.*` does too), `kinds`, and a `cursor` that takes either a `seq` or a `time_us`, told
+apart by magnitude. The index stores a `time_us`, since that also gives the cursor's age.
+`src/index/jetstream.ts` targets this v2 shape; `wrangler.jsonc`'s `JETSTREAM_URL` points at
+`wss://jetstream.us-east.bsky.network` accordingly.
+
+`com.atproto.sync.listReposByCollection` takes `collection`, `limit`, `cursor` and returns
+`{ repos: [{did}], cursor }`.
 
 ## Local dev
 
@@ -121,8 +112,4 @@ The Jetstream parser and ingest tests run on frames captured from the live v2 en
 writes `run.nibling.*` yet, the `run.nibling.*` fixtures keep a captured envelope and change only
 `collection` and `record`.
 
-A smoke run of `wrangler dev --local` confirmed `GET /oauth/client-metadata.json` (200, the exact
-D19 scope string) and `GET /xrpc/run.nibling.actor.getKnown?dids=did:plc:test` (200,
-`{"dids":[]}`, with CORS and cache headers) both serve correctly end to end.
-
 This worker is never deployed from a development session: no `wrangler deploy`, ever.
diff --git a/packages/api/src/index/jetstream.ts b/packages/api/src/index/jetstream.ts
index da33641..e9f51ce 100644
--- a/packages/api/src/index/jetstream.ts
+++ b/packages/api/src/index/jetstream.ts
@@ -10,15 +10,11 @@ function isIndexCollection(value: string): value is IndexCollection {
 }
 
 /**
- * Builds the live-tail subscribe URL.
- *
- * Deviation from the api-spec draft: the spec assumed Jetstream's legacy `/subscribe` endpoint
- * (`wantedCollections=...&cursor=<time_us>`). The current docs (bsky.network/docs/jetstream,
- * read 2026-11-06) describe a v2 API at `/xrpc/network.bsky.jetstream.subscribeEvents` with
- * repeated `collections` params, `kinds`, and a `cursor` that's either a `seq` or a `time_us`,
- * told apart by magnitude. The index passes a `time_us` (see ingest.ts). `collections` does document NSID-prefix wildcards
- * ("an `app.bsky.feed.*`-style namespace wildcard"), so `run.nibling.*` would also work, but
- * since there are only four collections to ask for, this lists them explicitly instead.
+ * Jetstream v2 is `/xrpc/network.bsky.jetstream.subscribeEvents` with repeated `collections`
+ * params, `kinds`, and a `cursor` that's either a `seq` or a `time_us`, told apart by magnitude.
+ * The index passes a `time_us` (see ingest.ts). `collections` supports NSID-prefix wildcards
+ * (an `app.bsky.feed.*`-style namespace wildcard), so `run.nibling.*` would also work, but since
+ * there are only four collections to ask for, this lists them explicitly instead.
  */
 export function subscribeUrl(baseUrl: string, cursor?: string): string {
   const url = new URL('/xrpc/network.bsky.jetstream.subscribeEvents', baseUrl);
diff --git a/packages/api/tests/backfill.test.ts b/packages/api/tests/backfill.test.ts
index ae0e0b1..82de17b 100644
--- a/packages/api/tests/backfill.test.ts
+++ b/packages/api/tests/backfill.test.ts
@@ -40,7 +40,6 @@ test('backfillCollection pages listReposByCollection and listRecords, and valida
               value: { $type: 'run.nibling.project', name: 'Alice Project', files: [], createdAt: '2026-11-01T00:00:00.000Z' },
             },
             {
-              // fails validation (no `name`): dropped, not thrown
               uri: 'at://did:plc:alice/run.nibling.project/3la2',
               cid: 'bafy2',
               value: { $type: 'run.nibling.project', files: [], createdAt: '2026-11-01T00:00:00.000Z' },
@@ -48,7 +47,6 @@ test('backfillCollection pages listReposByCollection and listRecords, and valida
           ],
         });
       }
-      // bob's PDS is unreachable
       throw new Error('network down');
     }
 
diff --git a/packages/api/tests/model.test.ts b/packages/api/tests/model.test.ts
index be1c2f3..059b074 100644
--- a/packages/api/tests/model.test.ts
+++ b/packages/api/tests/model.test.ts
@@ -127,8 +127,8 @@ test('delete removes the project from recent', () => {
 
 test('a like on a not-yet-indexed project still counts once the project appears', () => {
   const model = new IndexModel(sqliteSql());
-  model.apply(fixtureEvents()[1]!); // like1, subject = project1, before project1 is indexed
-  model.apply(fixtureEvents()[0]!); // project1 now appears
+  model.apply(fixtureEvents()[1]!);
+  model.apply(fixtureEvents()[0]!);
 
   const { projects } = model.recent();
   assert.equal(projects[0]!.likeCount, 1);
@@ -149,13 +149,7 @@ test('known returns exactly the DIDs with a project or profile record', () => {
   const model = new IndexModel(sqliteSql());
   for (const event of fixtureEvents()) model.apply(event);
 
-  const known = model.known([
-    'did:plc:alice', // has a project
-    'did:plc:bob', // only has a like
-    'did:plc:dave', // has a project
-    'did:plc:erin', // only has a profile
-    'did:plc:frank', // has nothing
-  ]);
+  const known = model.known(['did:plc:alice', 'did:plc:bob', 'did:plc:dave', 'did:plc:erin', 'did:plc:frank']);
 
   assert.deepStrictEqual(known, ['did:plc:alice', 'did:plc:dave', 'did:plc:erin']);
 });