@talqing/mcp 0.13.1 → 0.14.0

package/SKILL.md

@@ -376,18 +376,20 @@ states what each entry supports: which channels, its language codes and default
 language, whether it accepts a speed and within what range, and its default
 voice. Read it before choosing; do not assume a model exists.
 
-### Images
+### Images and files
 
-People can attach a photo to a web chat, a web voice call or a web video call,
-and the agent sees it. **There is no setting for this** — whether an agent can
-read images is decided by the model it runs, and `vision` on that model's catalog
-entry is where it says so. Most entries read images; `gemini-3.8-flash`,
-`gemini-3.7-flash`, `openai/gpt-live-1` and the two `grok-voice-*` realtime
-entries do not, and an agent on one of them cannot be sent a photo at all. If a
-user wants image input, check `vision` before picking the model rather than
-changing anything on the agent.
+People can send files on WhatsApp (Twilio senders) and in a web chat, and a photo
+on a web voice or video call. Every file is stored on the conversation; photos,
+PDFs and Word (`.docx`) files are also read by the agent, and anything else
+(voice notes, videos, spreadsheets) is named to it as something it cannot open.
+**There is no setting for this.** Whether the agent sees a photo or a scanned
+PDF page is decided by the model it runs: `vision` on its catalog entry. Most
+entries read images; `gemini-3.8-flash`, `gemini-3.7-flash`, `openai/gpt-live-1`
+and the two `grok-voice-*` realtime entries do not — such an agent is told it
+cannot view the image. Text PDFs and Word files are read on any model. If a user
+wants image input, check `vision` before picking the model.
 
-A phone call carries no files, so this is web only.
+A phone call carries no files.
 
 ### Vision input (screen share)
 
@@ -613,6 +615,9 @@ A tool is one capability the agent's LLM can call mid-conversation. It has:
   to get it out of the conversation, and a bare `{"type": "string"}` is how a
   tool ends up called with the wrong value. Mark the ones the tool cannot run
   without as `required`, and use an `enum` wherever the set of values is closed.
+  A **File** parameter is `{"type": "string", "format": "attachment"}`: the agent
+  passes the id of a file the person sent, and the operation reads
+  `{{args.<name>.url}}` (a link that lives an hour) and `.filename`, `.mime_type`.
 - **`long_running_task`** — voice and video agents keep talking while it runs and
   share the result when it lands; a run finishing within a second answers inline
   instead. Text agents wait either way. There is no fixed line — steer it from
@@ -1614,16 +1619,16 @@ connects and then fails.
 
 A chat is one text session, as a call is a voice one. `create_chat_message`
 takes a `client_message_id` (a UUID you generate) and returns the reply itself:
-`status`, `items` and `chat_status`. A chat **ends only explicitly** — the
-agent's `end_call`, or `end_chat` — and only then is it analysed and its
-`session.completed` sent, so end the chats you start. On Telegram and WhatsApp a
-person's messages join one chat until it ends; the next starts a new one.
-
-A message can carry `images`: up to four `{data_url, filename}` entries beside
-`message`, each a base64 `data:image/...` URL of a JPEG, PNG or WebP under 10 MB.
-The caption and the photo arrive as one turn and one item, so send them together
-rather than as two messages. The agent's model must have `vision` (see *Images*
-above) or the call is refused naming the model.
+`status`, `items` and `chat_status`. A chat ends with the agent's `end_call`,
+`end_chat`, or 24 hours after the person last wrote (`inactivity_timeout`) — only
+then is it analysed and its `session.completed` sent, so end the chats you start.
+On Telegram and WhatsApp a person's messages join one chat until it ends; the
+next starts a new one.
+
+A message can carry `files`: up to four `{data_url, filename}` entries beside
+`message`, each a base64 `data:` URL of any type up to 20 MB. The caption and the
+file arrive as one turn and one item, so send them together rather than as two
+messages. See *Images and files* above for what the agent reads.
 
 **Reading it back.** A *conversation* is the thread with a person, on any
 surface; a *call* is one voice or video session. A person is one **contact**

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@talqing/mcp",
-  "version": "0.13.1",
+  "version": "0.14.0",
   "description": "MCP server for building Talqing AI voice, video and text agents from Claude Code, Codex, or any MCP client.",
   "license": "MIT",
   "homepage": "https://talqing.com",

package/tools.json

@@ -3008,7 +3008,7 @@
       "role": "read"
     },
     {
-      "description": "One chat with its health snapshot, cost lines, vars and event trace.",
+      "description": "One chat with its health snapshot, cost lines, vars, what it ran and event trace.",
       "method": "GET",
       "name": "get_chat_detail",
       "parameters": {
@@ -3505,6 +3505,31 @@
       "read_only": true,
       "role": "read"
     },
+    {
+      "description": "One file a person sent in this conversation, with a fresh link (it lives 1 hour).",
+      "method": "GET",
+      "name": "get_conversation_attachment",
+      "parameters": {
+        "additionalProperties": false,
+        "description": "Arguments not inlined: call describe_function('get_conversation_attachment').",
+        "properties": {
+          "attachment_id": {
+            "type": "string"
+          },
+          "conversation_id": {
+            "type": "string"
+          }
+        },
+        "required": [
+          "conversation_id",
+          "attachment_id"
+        ],
+        "type": "object"
+      },
+      "path": "/v1/conversations/{conversation_id}/attachments/{attachment_id}",
+      "read_only": true,
+      "role": "read"
+    },
     {
       "description": "The agent sessions behind a conversation, newest first.",
       "method": "GET",