Update phrasing

.well-known/ai-plugin.json

@@ -3,7 +3,7 @@
   "name_for_model": "retrieval",
   "name_for_human": "Retrieval Plugin",
   "description_for_model": "Plugin for searching through the user's documents (such as files, emails, and more) to find answers to questions and retrieve relevant information. Use it whenever a user asks something that might be found in their personal information.",
-  "description_for_human": "Search through your documents",
+  "description_for_human": "Search through your documents.",
   "auth": {
     "type": "user_http",
     "authorization_type": "bearer"

.well-known/openapi.yaml

@@ -9,7 +9,7 @@ paths:
   /query:
     post:
       summary: Query
-      description: This endpoint accepts an array of search query objects, each containing a natural language query string ("query") and an optional metadata filter ("filter"). Filters can help refine search results based on criteria such as document source or time period, but are not necessary in most cases. You can send multiple queries to compare information from different sources or break down complex questions into sub-questions. If you receive a ResponseTooLargeError, try splitting up the queries into multiple calls to this endpoint.
+      description: Accepts search query objects array each with query and optional filter. Break down complex questions into sub-questions. Refine results by criteria, e.g. time / source, don't do this often. Split queries if ResponseTooLargeError occurs.
       operationId: query_query_post
       requestBody:
         content:

README.md

@@ -587,7 +587,7 @@ BEARER_TOKEN=your_bearer_token \
 
 Alternatively, you could set environment variables in the [Fly.io Console](https://fly.io/dashboard).
 
-At this point, you can change the plugin url in your plugin manifest file [here](/.well-known/ai-plugin.json), and in your OpenAPI schema [here](/.well-known/openapi.yaml) to the url for your Fly.io app, which will be https://<your-app-name>.fly.dev.
+At this point, you can change the plugin url in your plugin manifest file [here](/.well-known/ai-plugin.json), and in your OpenAPI schema [here](/.well-known/openapi.yaml) to the url for your Fly.io app, which will be `https://your-app-name.fly.dev`.
 
 Deploy your app with:
 
@@ -710,7 +710,7 @@ To view your app logs:
 heroku logs --tail -a [app-name]
 ```
 
-Now make sure to change the plugin url in your plugin manifest file [here](/.well-known/ai-plugin.json), and in your OpenAPI schema [here](/.well-known/openapi.yaml), and redeploy with `make heroku-push`. This url will be `https://<your-app-name>.herokuapp.com`.
+Now make sure to change the plugin url in your plugin manifest file [here](/.well-known/ai-plugin.json), and in your OpenAPI schema [here](/.well-known/openapi.yaml), and redeploy with `make heroku-push`. This url will be `https://your-app-name.herokuapp.com`.
 
 ### Other Deployment Options
 

examples/authentication-methods/no-auth/ai-plugin.json

@@ -3,7 +3,7 @@
   "name_for_model": "retrieval",
   "name_for_human": "Retrieval Plugin",
   "description_for_model": "Plugin for searching through the user's documents (such as files, emails, and more) to find answers to questions and retrieve relevant information. Use it whenever a user asks something that might be found in their personal information.",
-  "description_for_human": "Plugin to search through your personal documents.",
+  "description_for_human": "Search through your documents.",
   "auth": {
     "type": "none"
   },

examples/authentication-methods/no-auth/main.py

@@ -84,7 +84,7 @@ async def query_main(
 @sub_app.post(
     "/query",
     response_model=QueryResponse,
-    description='Accepts an array of search query objects, each with a natural language query string ("query") and an optional metadata filter ("filter"). Filters are not necessary in most cases, but can sometimes help refine search results based on criteria such as document source or time period. Send multiple queries to compare information from different sources or break down complex questions into sub-questions. If you receive a ResponseTooLargeError, try splitting up the queries into multiple calls to this endpoint.',
+    description="Accepts search query objects with query and optional filter. Break down complex questions into sub-questions. Refine results by criteria, e.g. time / source, don't do this often. Split queries if ResponseTooLargeError occurs.",
 )
 async def query(
     request: QueryRequest = Body(...),

examples/authentication-methods/oauth/ai-plugin.json

@@ -3,7 +3,7 @@
   "name_for_model": "retrieval",
   "name_for_human": "Retrieval Plugin",
   "description_for_model": "Plugin for searching through the user's documents (such as files, emails, and more) to find answers to questions and retrieve relevant information. Use it whenever a user asks something that might be found in their personal information.",
-  "description_for_human": "Plugin to search through your personal documents.",
+  "description_for_human": "Search through your documents.",
   "auth" : {
     "type":"oauth",
     "client_url":"e.g. https://<your domain>/oauth/v2/authorize",

examples/authentication-methods/service-http/ai-plugin.json

@@ -3,7 +3,7 @@
   "name_for_model": "retrieval",
   "name_for_human": "Retrieval Plugin",
   "description_for_model": "Plugin for searching through the user's documents (such as files, emails, and more) to find answers to questions and retrieve relevant information. Use it whenever a user asks something that might be found in their personal information.",
-  "description_for_human": "Plugin to search through your personal documents.",
+  "description_for_human": "Search through your documents.",
   "auth":{
 		"type":"service_http",
 		"authorization_type":"bearer",

examples/authentication-methods/user-http/ai-plugin.json

@@ -3,7 +3,7 @@
   "name_for_model": "retrieval",
   "name_for_human": "Retrieval Plugin",
   "description_for_model": "Plugin for searching through the user's documents (such as files, emails, and more) to find answers to questions and retrieve relevant information. Use it whenever a user asks something that might be found in their personal information.",
-  "description_for_human": "Plugin to search through your personal documents.",
+  "description_for_human": "Search through your documents.",
   "auth": {
     "type": "user_http",
     "authorization_type": "bearer"

examples/memory/ai-plugin.json

@@ -3,7 +3,7 @@
     "name_for_model": "retrieval",
     "name_for_human": "Retrieval Plugin",
     "description_for_model": "Plugin for searching through the user's documents (such as files, emails, and more) to find answers to questions and retrieve relevant information. Use it whenever a user asks something that might be found in their personal information, or asks you to save information for later.",
-    "description_for_human": "Search through your documents",
+    "description_for_human": "Search through your documents.",
     "auth": {
       "type": "user_http",
       "authorization_type": "bearer"

examples/memory/main.py

@@ -80,7 +80,8 @@ async def upsert_main(
 @sub_app.post(
     "/upsert",
     response_model=UpsertResponse,
-    description="Save information from chat conversations as documents, only if the user asks you to. Accepts an array of documents, each document has a text field with the conversation text and possible questions that could lead to the answer, and metadata including the source (chat) and created_at timestamp. Confirm with the user before saving information, and ask if they want to add details / context.",
+    # NOTE: We are describing the the shape of the API endpoint input due to a current limitation in parsing arrays of objects from OpenAPI schemas. This will not be necessary in future.
+    description="Save chat information. Accepts an array of documents with text (potential questions + conversation text), metadata (source 'chat' and timestamp). Confirm with the user before saving, ask for more details/context.",
 )
 async def upsert(
     request: UpsertRequest = Body(...),
@@ -115,7 +116,8 @@ async def query_main(
 @sub_app.post(
     "/query",
     response_model=QueryResponse,
-    description='Accepts an array of search query objects, each with a natural language query string ("query") and an optional metadata filter ("filter"). Filters are not necessary in most cases, but can sometimes help refine search results based on criteria such as document source or time period. Send multiple queries to compare information from different sources or break down complex questions into sub-questions. If you receive a ResponseTooLargeError, try splitting up the queries into multiple calls to this endpoint.',
+    # NOTE: We are describing the the shape of the API endpoint input due to a current limitation in parsing arrays of objects from OpenAPI schemas. This will not be necessary in future.
+    description="Accepts search query objects array each with query and optional filter. Break down complex questions into sub-questions. Refine results by criteria, e.g. time / source, don't do this often. Split queries if ResponseTooLargeError occurs.",
 )
 async def query(
     request: QueryRequest = Body(...),

examples/memory/openapi.yaml

@@ -9,7 +9,7 @@ paths:
   /upsert:
     post:
       summary: Upsert
-      description: Save information from chat conversations as documents, only if the user asks you to. Accepts an array of documents, each document has a text field with the conversation text and possible questions that could lead to the answer, and metadata including the source (chat) and created_at timestamp. Confirm with the user before saving information, and ask if they want to add details / context.
+      description: Save chat information. Accepts an array of documents with text (potential questions + conversation text), metadata (source 'chat' and timestamp). Confirm with the user before saving, ask for more details/context.
       operationId: upsert_upsert_post
       requestBody:
         content:
@@ -35,7 +35,7 @@ paths:
   /query:
     post:
       summary: Query
-      description: Accepts an array of search query objects, each with a natural language query string ("query") and an optional metadata filter ("filter"). Filters are not necessary in most cases, but can sometimes help refine search results based on criteria such as document source or time period. Send multiple queries to compare information from different sources or break down complex questions into sub-questions. If you receive a ResponseTooLargeError, try splitting up the queries into multiple calls to this endpoint.
+      description: Accepts search query objects array each with query and optional filter. Break down complex questions into sub-questions. Refine results by criteria, e.g. time / source, don't do this often. Split queries if ResponseTooLargeError occurs.
       operationId: query_query_post
       requestBody:
         content:

server/main.py

@@ -94,7 +94,8 @@ async def query_main(
 @sub_app.post(
     "/query",
     response_model=QueryResponse,
-    description='Accepts an array of search query objects, each with a natural language query string ("query") and an optional metadata filter ("filter"). Filters are not necessary in most cases, but can sometimes help refine search results based on criteria such as document source or time period. Send multiple queries to compare information from different sources or break down complex questions into sub-questions. If you receive a ResponseTooLargeError, try splitting up the queries into multiple calls to this endpoint.',
+    # NOTE: We are describing the the shape of the API endpoint input due to a current limitation in parsing arrays of objects from OpenAPI schemas. This will not be necessary in future.
+    description="Accepts search query objects array each with query and optional filter. Break down complex questions into sub-questions. Refine results by criteria, e.g. time / source, don't do this often. Split queries if ResponseTooLargeError occurs.",
 )
 async def query(
     request: QueryRequest = Body(...),