Changelog

0.26.0 — 2026-06-16

First public release as grammy-testing

This is the first public npm release. The package is published under the third-party name grammy-testing (previously scaffolded as @grammyjs/testing), following the grammY third-party plugin convention (grammy-<name>). Install with npm install --save-dev grammy-testing and import from grammy-testing / grammy-testing/low-level. JSR publishing is deferred for now; the @grammyjs/testing name is reserved for a future official grammY release.

grammY 1.44 / Bot API 10.0 + 10.1 support

Bumps the grammy peer and dev dependency to ^1.44.0 (Bot API 10.1, @grammyjs/types@3.28.0) and adds first-class support for the new surfaces. All shapes were verified against the published types.

• New message-sending methods: sendRichMessage and sendLivePhoto now return a synthetic Message and are routed into chat.messages / user.replies, like other senders.
• Rich message accessor: reply.richMessage exposes the sent InputRichMessage (html / markdown / is_rtl / skip_entity_detection) with a plainText convenience.
• Message drafts: sendMessageDraft / sendRichMessageDraft (both return true) are captured into a drafts projection — user.drafts and chats.draftsFor(user) — for streaming-sequence assertions.
• Guest mode: user.sendGuestMessage(chat, text?) dispatches a guest_message update and returns the generated guest_query_id. answerGuestQuery resolves with a synthetic SentGuestMessage ({ inline_message_id }) and is correlated to the originating user via chats.guestQueryUser(queryId); it is not routed into chat.messages.
• Join-request queries: user.requestJoin(group) now emits and returns the chat_join_request.query_id.
• Reaction removal: deleteMessageReaction / deleteAllMessageReactions are captured into chats.reactionRemovals.
• Managed-bot defaults: synthetic defaults for getManagedBotAccessSettings, setManagedBotAccessSettings, getManagedBotToken, replaceManagedBotToken, and getUserPersonalChatMessages; getChatAdministrators honors return_bots: false.
• Examples & docs: new 24-guest-mode-bot, 25-rich-message-bot, and 26-reaction-removal-bot examples plus a "Bot API 10 Features" documentation page.

0.25.0 — 2026-05-07

Transformer infrastructure

• TerminalTransformer type: createTransformer in src/low-level/transformer.ts now returns an internal TerminalTransformer type whose signature omits _previous. A new asTransformer adapter in prepare-bot.ts converts it for bot.api.config.use. The 4-line prose comment explaining why _previous was never called is replaced by this compile-time invariant.
• respondNextRaw(method, response): New method on OutgoingRequests (accessible as chats.outgoing.respondNextRaw) that injects a verbatim raw response for the next matching API call — bypassing the { ok: true, result } wrapper. Use it to simulate rate-limit responses ({ ok: false, error_code: 429, parameters: { retry_after: 0 } }) that outer transformers such as @grammyjs/auto-retry can observe and act on.
• Auto-retry retry-on-429 test: tests/plugins/auto-retry.spec.ts now includes a test that verifies autoRetry retries a sendMessage call when respondNextRaw injects a 429 raw response. Two sendMessage entries appear in chats.outgoing.requests (original + retry).

0.24.1 — 2026-05-07

Code quality

• createTransformer terminal-intent comment: _previous in src/low-level/transformer.ts is now annotated with an inline comment explaining it is intentionally never called. This documents the invariant that the snapshot-and-reinstall pattern in prepareBot relies on.
• Plugin example context flavor types: examples/21-files-bot now uses FileFlavor<Context> and examples/22-hydrate-bot uses HydrateFlavor<Context>, replacing as unknown as casts with proper plugin-exported flavor types.
• Test helper cleanup: Empty /** */ JSDoc blocks removed from the four private helper functions in tests/plugins/chat-members.spec.ts.
• ESLint examples alignment: examples/**/*.ts JSDoc rules are no longer silenced — examples are now held to the same JSDoc standards as src/.

0.24.0 — 2026-05-07

Plugin interop: grammy-media-groups

mediaGroupTransformer(adapter) installed via bot.api.config.use() before prepareBot now runs correctly during tests and stores outgoing media group messages in the adapter.

• syntheticMediaGroup response shape: every message in the default sendMediaGroup response now includes a chat field (required by storeMessages for deduplication by chat.id) and a media_group_id string shared across all messages in the same call (required for adapter grouping). This is a minor additive change — bots reading sendMediaGroup return values in tests will see the additional fields.
• grammy-media-groups added to plugin interop table: install mediaGroupTransformer(adapter) before prepareBot, assert on adapter.read(media_group_id).

See tests/plugins/media-groups.spec.ts and site/plugins/media-groups.md.

0.23.0 — 2026-05-07

Plugin transformer support

Bot-level transformers installed via bot.api.config.use() are now correctly chained during tests. Previously the library's mock transformer was installed last (outermost), silently skipping all user-installed transformers. With this fix, the library transformer is always innermost — every transformer you install runs normally and can process synthetic responses.

• Transformer chain fix: prepareBot now snapshots existing transformers, installs the library transformer first, then reinstalls user transformers on top. Uses only the public installedTransformers() / use() API — no private access.
• Realistic getFile default: buildDefaultResponses now returns a valid File shape for getFile (file_id, file_unique_id, file_size, file_path) instead of true, so @grammyjs/files can hydrate it without a custom responses override.
• Synthetic message includes chat: Default sendMessage / sendPhoto / etc. responses now include a chat field so @grammyjs/hydrate can attach delete() / edit() methods.

New plugin interop: @grammyjs/files

hydrateFiles(bot.token) installed before prepareBot runs correctly. ctx.getFile() returns a hydrated File with getUrl() and download() methods.

See tests/plugins/files.spec.ts and examples/21-files-bot/.

New plugin interop: @grammyjs/hydrate

hydrateApi() (bot-level transformer) and hydrate() (context middleware) both work. Bot API call results include delete() / edit() / pin() convenience methods. Context objects get ctx.message.delete() and similar shortcuts.

See tests/plugins/hydrate.spec.ts and examples/22-hydrate-bot/.

New plugin interop: @grammyjs/auto-retry

autoRetry(options) installed before prepareBot is now part of the transformer chain for every API call. Normal bot operation is unaffected. failNext errors propagate through autoRetry to the handler catch block (autoRetry does not retry thrown GrammyErrors).

See tests/plugins/auto-retry.spec.ts and examples/23-auto-retry-bot/.

New plugin interop: @grammyjs/chat-members — hydrateChatMember()

hydrateChatMember() API transformer is now supported. Install via bot.api.config.use(hydrateChatMember()) before prepareBot. getChatMember and getChatAdministrators results are augmented with an .is(query) method at test time, matching production behaviour.

See tests/plugins/chat-members.spec.ts (new hydrateChatMember() describe block).

Custom transformer chain support

Request-mutating and response-augmenting transformers installed via bot.api.config.use() before prepareBot now run normally. Payload mutations are visible in chats.outgoing.requests; response augmentations are visible to handlers.

See tests/plugins/custom-transformer.spec.ts.

VitePress: new Plugins section

A dedicated Plugins sidebar group replaces the mixed plugin/recipe content in Recipes:

• conversations and menu pages moved from site/recipes/ to site/plugins/
• Five new pages: Chat Members, Files, Hydrate, Auto-Retry, Transformer Throttler
• Recipes retains only general-pattern pages: Sessions, Keyboards, Error Simulation, Multi-Chat Scenarios, Fire & Forget

0.22.0 — 2026-05-05

VitePress documentation site

Standalone documentation site under site/ — 42 pages covering every public export.

• Infrastructure: site/.vitepress/config.ts with dynamic GitHub Pages base URL, full nav/sidebar, local search, social links, and version badge from package.json.
• Logo: docs/Y.svg (the grammY Y mark) copied to site/public/logo.svg.
• Brand theme: blue palette — #0057b7 in light mode (6.9:1 on white, WCAG AA) / #4d9eff in dark mode (6.3:1 on dark background, WCAG AA).
• Guide section (6 pages): Introduction, Getting Started, How It Works, With Vitest, With Jest, With Deno.
• High-Level API section (9 pages): Chats, User, Group & Supergroup, Channel, PrivateChat, BusinessAccount, Reply, Logs, Overview.
• Low-Level API section (5 pages): Outgoing Requests, Session Mocking, Update Builders, Response Mocking, Overview.
• Recipes section (7 pages): Sessions & State, Keyboards & Buttons, Error Simulation, Multi-Chat Scenarios, Conversations Plugin, Menu Plugin, Fire & Forget.
• API Reference section (14 pages): full typed signatures for every exported symbol.
• Reference: Changelog page linked to docs/CHANGELOG.md.
• GitHub Actions workflow (.github/workflows/docs.yml): builds and deploys to GitHub Pages at https://drsmile444.github.io/grammy-testing/ on every push to main.

0.21.0 — 2026-05-05

README rewrite

README.md has been completely replaced. The previous file was a TypeScript Boilerplate placeholder with no relation to the library. The new README is production-grade and client-facing:

• Logo (docs/Y.svg — the grammY Y mark) with yellow #ffd700 brand color across all badges
• Why section with bold hook sentence and problem/solution statement
• Quick Start with npm + JSR install commands and a /start command bot example
• Features section covering high-level actors, dispatch verbs, assertions, session injection, and isolation utilities
• Examples table linking all 20 examples in examples/
• Documentation placeholder pointing to the planned VitePress site
• Credits section acknowledging grammy_tests and ua-anti-spam-bot

0.20.0 — 2026-05-03

20 runnable examples added under examples/

A new examples/ directory ships 20 self-contained bots with matching test files, covering a wide range of testing patterns:

#	Scenario
01	Echo bot — simplest possible text-echo handler
02	Start command — /start reply
03	Greeting bot — per-user name with fallback
04	Chat-type filter — private vs. group routing
05	Regex handler — pattern-matched messages
06	Callback query — inline keyboard responses
07	Session counter — persistent per-user state
08	Chat settings — mockChatSession usage
09	Photo bot — caption extraction
10	Document bot — file-ID and MIME type reply
11	Poll bot — quiz creation and answer scoring
12	Group welcome — new_chat_members service event
13	Admin guard — getChatMember status check
14	Moderation bot — banChatMember / restrictChatMember
15	Channel post bot — channel_post handler
16	Reactions bot — message_reaction handler
17	Dice game — incoming dice value evaluation
18	Middleware test — prepareMiddleware isolation
19	Composer test — prepareComposer isolation
20	Multi-chat scenario — cross-chat summary posting

All examples are included in the test run and pass the full quality gate.

0.19.0 — 2026-05-03

user.sendCallbackQuery and clickButton reply_markup fix

user.sendCallbackQuery(data, options?) is now available. Dispatches a callback_query update from the user without requiring a prior captured reply. Useful for cross-feature tests where the keyboard lives in a different composer.

await user.sendCallbackQuery('button-data');

// With explicit message context (for chatType filters or handlers that read keyboard state):
const msg = await user.sendCommand('/menu');
await user.sendCallbackQuery('language:en', { message: msg });

When options.message is omitted, a minimal private-chat stub is synthesized automatically so grammY filters like chatType('private') evaluate correctly without boilerplate.

clickButton now populates callback_query.message.reply_markup. Reply.toCapturedMessage() previously omitted reply_markup, leaving ctx.callbackQuery.message.reply_markup as undefined in the handler. Handlers that read keyboard state (e.g. ctx.callbackQuery.message.reply_markup.inline_keyboard) now receive the full keyboard.

0.18.0 — 2026-05-03

Internal fixes: update_id counter independence, sendMediaGroup response shape, GROUP_ANONYMOUS_BOT documentation

update_id counter is now independent from message IDs. All User dispatch methods (sendPhoto, sendDocument, sendVideo, sendAudio, sendVoice, sendVideoNote, sendAnimation, sendSticker, sendLocation, sendContact, sendVenue, sendPoll, sendDice, sendWebAppData, sendSuccessfulPayment, sendInlineQuery, sendChosenInlineResult, sendPreCheckoutQuery, sendShippingQuery, reactTo, answerPoll, requestJoin, boostChat, removeBoost, manageBot, purchasePaidMedia, sendMediaGroup) and Reply.clickButton now draw update_id from IdGenerator.nextUpdateId() rather than nextMessageId() + offset. Message IDs and update IDs no longer share a counter.

sendMediaGroup auto-response returns N messages. The default buildDefaultResponses resolver for sendMediaGroup now returns an array whose length matches the number of items in the bot's media payload. Previously it always returned a single-element array regardless of how many media items were sent.

GROUP_ANONYMOUS_BOT.is_bot: false is intentional. An inline comment documents that Telegram sends is_bot: false for this identity in real update payloads, consistent with Channel_Bot (id: 136 817 688).

0.17.0 — 2026-05-03

channel.postMessageTo returns Message and accepts reply_to_message

channel.postMessageTo(target, text, options?) now returns Promise<Message> (previously Promise<void>), consistent with all other send verbs since v0.15.0. The returned value can be used immediately as reply_to_message in a follow-up user.sendText:

const post = await channel.postMessageTo(group, 'announcement');
await user.sendText('nice post', { chat: group, reply_to_message: post });

A new reply_to_message option is also accepted. It follows the same partial-shape semantics as user.sendText: date and chat are auto-filled when absent, and all explicitly supplied fields are preserved:

await channel.postMessageTo(group, 'reply', {
  reply_to_message: { message_id: 10 },
  // date and chat are auto-filled from context
});

0.16.0 — 2026-05-03

Relay message support (group.postRelayMessage, TELEGRAM_RELAY)

Groups and supergroups now have a dedicated postRelayMessage verb that dispatches the synthetic message update Telegram produces when a channel post is forwarded into a linked group (from.id === 777_000). The returned Message can be passed directly as reply_to_message in a follow-up user.sendText:

const relay = await group.postRelayMessage('channel post');
await user.sendText('my comment', { chat: group, reply_to_message: relay });

To simulate a relayed post with channel attribution, pass options.channel:

const channel = chats.newChannel('My Channel');
const relay = await group.postRelayMessage('post text', { channel });
// ctx.message.forward_origin.type === 'channel'

A TELEGRAM_RELAY constant is exported for assertions:

import { TELEGRAM_RELAY } from 'grammy-testing';
expect(ctx.message.from).toMatchObject(TELEGRAM_RELAY);

Partial reply_to_message in SendTextOptions

SendTextOptions.reply_to_message now accepts Partial<Message> & { message_id: number } — only message_id is required. date and chat are auto-filled when absent:

// No more `as any` casts or manual date/chat construction
await user.sendText('reply', { chat: group, reply_to_message: { message_id: 42 } });

Callers that already pass a full Message are unaffected.

0.15.0 — 2026-05-03

Actor send verbs return the dispatched Message

All User send verbs that produce a message update now return Promise<Message> instead of Promise<void>. The returned object is the exact synthetic Message dispatched to bot.handleUpdate, giving tests direct access to message_id, chat, from, and content fields without magic numbers or private state access:

const msg = await user.sendText('not a card');
await user.editMessage(msg.message_id, '4111 1111 1111 1111');
// or chain with reply_to_message:
await user.sendText('reply', { chat: group, reply_to_message: msg });

user.sendMediaGroup(items) returns Promise<Message[]> — one Message per dispatched item in order, all sharing the same media_group_id:

const [first, second] = await user.sendMediaGroup([{ photo: 'a' }, { photo: 'b' }]);
expect(first.media_group_id).toBe(second.media_group_id);

Existing callers that ignore the return value are unaffected — the change is fully backward-compatible.

Affected verbs: sendText, sendMessage, sendCommand, sendForwarded, sendPhoto, sendDocument, sendVideo, sendAudio, sendVoice, sendVideoNote, sendAnimation, sendSticker, sendLocation, sendContact, sendVenue, sendPoll, sendDice, sendWebAppData, sendSuccessfulPayment, sendMediaGroup.

---

0.14.0 — 2026-05-03

ChatProfile — chat factory methods accept a caller-supplied ID

chats.newGroup, chats.newSupergroup, and chats.newChannel now accept an optional object profile { id?, title? } in addition to the existing string/undefined forms, mirroring the UserProfile pattern used by chats.newUser. Any integer ID is accepted without validation — use this to register chats whose IDs are fixed at configuration time (log channels, training chats, etc.):

const logsGroup = chats.newSupergroup({ id: 1_234_567, title: 'Logs' });
// title defaults to 'Supergroup1234567' when omitted
const alerts = chats.newChannel({ id: -500, title: 'Alerts' });

getChat and getChatAdministrators auto-derivation work normally for specific-ID chats, eliminating the need for respondNext('getChat', ...) workarounds.

anonymous option — sendText / sendCommand dispatch as GroupAnonymousBot

SendTextOptions gains anonymous?: boolean. When true, the dispatched message's from is replaced with the GroupAnonymousBot identity and sender_chat is set to the target group, matching Telegram's wire format for the "Send as Group" admin feature:

await user.sendText('/role user', { chat: group, anonymous: true });
await user.sendCommand('/role', 'user', { chat: group, anonymous: true });

Requires options.chat to be a Group or Supergroup; throws a descriptive error otherwise.

GROUP_ANONYMOUS_BOT — exported constant for assertions

import { GROUP_ANONYMOUS_BOT } from 'grammy-testing';
// { id: 1_087_968_824, username: 'GroupAnonymousBot', is_bot: false, first_name: 'Group' }
expect(ctx.message.from).toMatchObject(GROUP_ANONYMOUS_BOT);

sendSystemMessage — dispatch a from-absent message update

Group, Supergroup, and Channel each gain sendSystemMessage(text, options?) which dispatches a message update with the from field intentionally absent. Tests the common if (!ctx.from) return next() guard path without raw handleUpdate calls:

await group.sendSystemMessage('no sender text');
await channel.sendSystemMessage('notice', { messageId: 42 });

---

0.13.0 — 2026-05-02

Channel.changeMemberStatus — dispatch my_chat_member for channels

• Added channel.changeMemberStatus(fromUser, transition) to Channel. Dispatches a my_chat_member update with chat.type === 'channel', updates the bot's membership in channel.members, and enables getChatAdministrators auto-derivation for channels. This closes the last remaining raw handleUpdate gap — all my_chat_member scenarios across Group, Supergroup, and Channel now use the same actor-verb API.
• Added CHANNEL_ADMIN_RIGHTS constant providing channel-appropriate defaults (can_post_messages: true; excludes can_manage_video_chats and can_manage_topics). Permissions supplied in the transition override the defaults.

Fix: changeMemberStatus now correctly tracks the bot's membership

Breaking (minor): group.changeMemberStatus(user, transition) and supergroup.changeMemberStatus(user, transition) previously stored the trigger actor (user) in the chat's members map and used that same user for old/new_chat_member.user in the dispatched update. Both were wrong — my_chat_member always describes the bot's status change, and from is the actor who triggered it.

After this fix:

• old/new_chat_member.user in the dispatched update is bot.botInfo (the bot), not the trigger user.
• The bot's membership is stored in the members map, keyed by bot.botInfo.id.
• getChatAdministrators auto-derivation now returns the bot after a promotion transition, not the trigger actor.
• The trigger user's own membership entry is not affected by changeMemberStatus.

Tests that called user.in(group) after changeMemberStatus to verify the new status should switch to group.members.get(bot.botInfo.id)?.status.

---

0.12.0 — 2026-05-01

group.own() and group.join() — pure-state membership setters

• Added group.own(user) and supergroup.own(user) — sets status: 'creator' in the members map with no Telegram update dispatched. Mirrors the existing promote() / restrict() pattern.
• Added group.join(user) and supergroup.join(user) — sets status: 'member' in the members map with no Telegram update dispatched.
• Added chats.newOwner(profile?) — convenience factory that creates a new user and calls defaultGroup.own(user), mirroring chats.newAdmin().

Auto-derived getChatMember, getChatAdministrators, getChat

• getChatMember now resolves from the registered chat's members map via a Membership → ChatMember converter. Returns the appropriate discriminated union shape ('creator', 'administrator', 'member', 'restricted', 'left', 'kicked'). Falls back to { status: 'left', user } for users not in the map, and to true for unregistered chats.
• getChatAdministrators now resolves by filtering chat.members for 'creator' and 'administrator' entries. Returns [] for unregistered chats.
• getChat now resolves from chat.toTelegramChat() enriched with invite_link: ''. Returns true for unregistered chats.
• All three are populated automatically in buildDefaultResponses(). User-supplied responses entries always take precedence — existing overrides are unaffected.

---

0.11.0 — 2026-05-01

chats.clear() — single-call state reset

• Added chats.clear() method that atomically resets all captured state: outgoing requests, per-user replies, actions, and edits logs, per-chat messages and deletions logs, and internal routing registries (messageIdToReply, clickers). User/chat registries and membership state are preserved, so existing user and group references remain valid. Replaces the previous 4–5 individual clear() calls required in beforeEach blocks.

warnOnUnregisteredChats — developer warning for silent misses

• Bot calls to sendMessage, sendPhoto, and other message-sending methods, sendChatAction, and deleteMessage targeting a chat ID not registered with the Chats orchestrator now emit a console.warn by default. The warning includes the method name, the unregistered chat ID, and guidance on how to register the chat or suppress the warning.
• Pass { warnOnUnregisteredChats: false } to prepareBot, prepareComposer, or prepareMiddleware to suppress the warning (useful for bots that intentionally fan out to external log channels).

Fix: postinstall script no longer breaks consumer installs

• Removed the postinstall entry from package.json. The ./scripts/link-codex-skills.sh hook is a local development convenience and is not present in the published package. Consumers no longer need npm install --ignore-scripts to work around the missing script error.

---

0.10.0 — 2026-05-01

Deletion tracking

• Added DeletionsLog per-chat log of deleteMessage calls, accessible via chats.deletionsFor(chat)
• Each deletion entry carries the synthetic message_id and a back-reference to the original Reply object (if the message was sent during the test)
• Exported new Deletion type

copyMessage and forwardMessage fixes

• copyMessage now returns a synthetic MessageId ({ message_id }) instead of true
• forwardMessage now returns a synthetic Message ({ message_id, date }) instead of true
• Both methods are now tracked by the Chats pipeline and produce Reply objects that appear in chat.messages and user.replies

---

0.9.0 — 2026-05-01

Synthetic Message responses

• All message-sending methods (sendMessage, sendPhoto, sendDocument, sendVideo, sendAudio, sendVoice, sendVideoNote, sendAnimation, sendSticker, sendLocation, sendContact, sendVenue, sendPoll, sendDice, sendMediaGroup) now return a real Message (or Message[] for sendMediaGroup) by default, using the synthetic message_id already assigned to the captured Reply
• User-supplied responses entries continue to override the defaults

State injection

• PrepareOptions gains an optional state field
• When provided, a mockState middleware is automatically inserted before the bot/composer under test so ctx.state is pre-populated for every dispatched update
• Compatible with prepareBot, prepareComposer, and prepareMiddleware

---

0.8.0 — 2026-05-01

User DX enhancements

• Added user.replies getter returning the user's RepliesInbox directly — no more chats.repliesFor(user) at every assertion site
• Added RepliesInbox.lastOrThrow() returning Reply<TContext> (non-nullable), throwing with a descriptive message when the inbox is empty
• Added chats.actionsFor(user) returning an ActionsLog that captures sendChatAction payloads for that user
• Added chats.editsFor(user) returning an EditsLog that captures editMessageText, editMessageCaption, and editMessageMedia calls resolved to that user

ID and counter fixes

• Fixed joinChat/leaveChat using hardcoded constants for update_id; all user actor dispatches now use nextUpdateId()
• Fixed sendText, sendForwarded, and editMessage deriving updateId from nextMessageId() instead of nextUpdateId()
• Implemented IdGenerator-scoped message IDs; removed all module-level counters to eliminate counter bleed between test runs in the same process

JSDoc coverage

• Added JSDoc to all public and non-trivial internal methods and constructors across src/
• Enabled jsdoc/require-jsdoc for class methods and constructors in ESLint config

CI improvements

• Dropped Node 18 from the test matrix; raised minimum engine to >=20.0.0
• Fixed verify-cjs.cjs require paths to resolve from project root
• Fixed npm/corepack CI workflow issues; switched to direct npm install

---

0.7.2 — 2026-04-30

Fixes and internal improvements

• Fixed recordClick not including chat_id in the callback routing record, causing button clicks in one chat to route replies globally
• Added nextUpdateId() method to IdGenerator
• Refactored OutgoingRequests.requests from a public mutable field to a read-only getter backed by a private array
• Extracted CJS verification to scripts/verify-cjs.cjs; lowered Node.js engine requirement to >=18.0.0
• Raised OutgoingRequests.getAll() typed overloads from 6 to 10 type parameters

---

0.7.1 — 2026-04-30

Type safety improvements

• Replaced hand-copied ParseMode union with a re-export from grammy, keeping it in sync with upstream automatically
• Converted MediaType union to a derived (typeof MEDIA_FIELDS)[number] type so adding a new media field is a compile-time error if the union is not updated
• Added exhaustiveness guard to makeChatMember switch so new grammy ChatMemberStatus variants produce a TypeScript error rather than silently falling through to 'kicked'

ESLint compliance

• Removed Plugin source overrides and Test overrides blocks from eslint.config.mjs; all violations in src/ and tests/ have been fixed instead
• Zero ESLint overrides in source and test code

---

0.7.0 — 2026-04-30

Business account API

• Added BusinessAccount high-level actor with verbs for all Telegram Business API update types:
  • connect(options?) → business_connection (is_enabled: true)
  • disconnect(options?) → business_connection (is_enabled: false)
  • sendMessage(text, options?) → business_message
  • editMessage(messageId, newText, options?) → edited_business_message
  • deleteMessages(messageIds, options?) → deleted_business_messages
• Added user.manageBot(botUser, options?) → managed_bot

Previously excluded update types

• Added user.purchasePaidMedia(payload, options?) → purchased_paid_media
• Added chat.dispatchReactionCount(messageId, reactions, options?) on Group, Supergroup, and Channel → message_reaction_count
• Added chats.dispatchPollState(poll, options?) → poll
• Removed all newly-covered types from the "Not covered" section in README

---

0.6.0 — 2026-04-30

Modern update types

• Added user.reactTo(reply, reaction) → message_reaction
• Added user.answerPoll(reply, optionIndices) → poll_answer
• Added user.requestJoin(group) → chat_join_request
• Added group.dispatchMemberUpdate(adminUser, targetUser, newStatus, options?) on Group and Supergroup → chat_member
• Added channel.editPost(messageId, newText, options?) → edited_channel_post
• Added user.boostChat(chat) → chat_boost
• Added user.removeBoost(chat, boostId) → removed_chat_boost

---

0.5.1 — 2026-04-29

Reply accessors

• Added reply.replyingTo — the earlier Reply object that this message is replying to
• Added reply.replyMarkup — raw reply_markup escape hatch for inspecting non-inline-keyboard markup

Private chat message log

• PrivateChat now exposes a messages log consistent with Group and Supergroup, capturing every bot message sent to a private DM

Context constructor option

• prepareComposer and prepareMiddleware now accept a ContextConstructor option in PrepareOptions, enabling bots with class-based custom context types to instantiate the correct runtime class

---

0.4.0 — 2026-04-29

Special message verbs

• Added user.sendInlineQuery(query, options?) → inline_query
• Added user.chooseInlineResult(resultId, options?) → chosen_inline_result
• Added user.sendWebAppData(data, buttonText, options?) → web_app_data
• Added user.completePurchase(options?) → successful_payment
• Added user.sendPreCheckoutQuery(options?) → pre_checkout_query
• Added user.sendShippingQuery(options?) → shipping_query

---

0.3.0 — 2026-04-28

Remaining dispatch verbs

• Added user.sendAudio(options?) → audio message
• Added user.sendVoice(options?) → voice message
• Added user.sendVideoNote(options?) → video_note message
• Added user.sendAnimation(options?) → animation message
• Added user.sendSticker(options?) → sticker message
• Added user.sendLocation(options?) → location message
• Added user.sendContact(options?) → contact message
• Added user.sendVenue(options?) → venue message
• Added user.sendPoll(question, options?, options2?) → poll message
• Added user.sendDice(options?) → dice message

---

0.2.0 — 2026-04-28

Media send verbs

• Added user.sendPhoto(options?) → single photo message
• Added user.sendDocument(options?) → document message
• Added user.sendVideo(options?) → video message
• Added user.sendMediaGroup(items) → dispatches N updates sharing a media_group_id, each with realistic file_id fields

---

0.1.1 — 2026-04-28

Plugin interop

• Added tests/plugins/ reference suite demonstrating grammy-testing usage alongside:
  • @grammyjs/conversations — multi-step conversation flows
  • @grammyjs/menu — inline menu navigation and callback routing
  • @grammyjs/parse-mode — ctx.replyWithHTML() / ctx.replyFmt() and parseMode assertions
  • @grammyjs/hydrate — hydrated message objects
  • @grammyjs/chat-members — member status tracking

---

0.1.0 — 2026-04-27

Low-level testing primitives

• Added prepareBot(bot, options?) entry point — sets up an in-process grammY bot with a captured transformer and returns { chats, bot }
• Added prepareComposer(composer, options?) and prepareMiddleware(middleware, options?) entry points for testing composers and middleware in isolation
• Added error simulation via options.responses — supply per-method canned responses or a function to produce them
• Added OutgoingRequests capture surface (outgoing.requests, outgoing.getAll(), outgoing.getLast(), outgoing.idle())
• Added low-level update builders: buildTextMessage, buildCallbackQuery, buildInlineQuery, buildMyChatMember, and others

High-level Chats/User/Admin API

• Added Chats<TContext> orchestrator with newUser(profile?), newGroup(name?), newSupergroup(name?), newChannel(name?), newPrivateChat(user) factory methods
• Added User<TContext> actor as the primary test subject driver
• Added Admin role via group.promote(user, perms?) and group.restrict(user, perms?)
• Added Reply<TContext> normalized object with text, parseMode, entities, buttons, replyMarkup, chat, replyingTo, raw, and clickButton(textOrCallbackData) for synthesizing callback queries
• Added user.replies and chat.messages inbox/log as the primary assertion surfaces
• Added chats.outgoing for raw outgoing request inspection

User dispatch verbs

• Added user.sendText(text, options?) and user.sendMessage(text, options?) (alias)
• Added user.sendCommand(command, args?, options?) with auto-emitted bot_command entity and optional chat target for group commands
• Added user.sendForwarded(reply, options?) and user.editMessage(reply, newText, options?)
• Added user.joinChat(chat, options?) and user.leaveChat(chat, options?) service-message verbs
• Added channel.postMessageTo(group, text, options?) for channel-as-author (sender_chat) scenarios

Build and CI

• Configured dual ESM + CJS exports with subpath export map (./, ./low-level, ./high-level)
• Added GitHub Actions CI matrix covering Node 20 and Node 22, CJS verification, and Bun
• Added jsr.json scaffold for future Deno/JSR publishing