> ## Documentation Index
> Fetch the complete documentation index at: https://docs.octen.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Broad Search
> Automatically decomposes user messages into multiple sub-queries, performs searches, and synthesizes results using an LLM.
## OpenAPI
````yaml /api-reference/openapi.json post /broad-search
openapi: 3.1.0
info:
title: Octen API
description: >-
Octen API provides Search, Extract, Embeddings, VL Embeddings, Web Chat,
Broad Search, and Deep Research services. The Search API searches ranked web
results with optional filters, highlights, and full content. The Extract API
extracts clean markdown content from URLs, with optional query-focused
highlights, page classification, and multimedia resources. The Embeddings
API converts text into vector representations. The VL Embeddings API
converts multimodal inputs (text, images, videos) into vector
representations. The Web Chat API provides LLM chat completions with search
augmentation. The Broad Search API automatically decomposes queries into
multiple sub-queries for comprehensive search and synthesis. The Deep
Research API runs a multi-round adaptive research pipeline that produces a
structured research plan, executes iterative web searches, builds a report
brief with evidence, and streams a final long-form report.
version: 1.0.0
servers:
- url: https://api.octen.ai
security:
- apiKeyAuth: []
paths:
/broad-search:
post:
summary: Broad Search
description: >-
Automatically decomposes user messages into multiple sub-queries,
performs searches, and synthesizes results using an LLM.
operationId: broad-search
requestBody:
required: true
content:
application/json:
schema:
$ref: '#/components/schemas/BroadSearchRequest'
examples:
basic:
summary: Basic Broad Search
value:
model: anthropic/claude-sonnet-4.6
messages:
- role: user
content: >-
Latest trends and major players in the global AI chip
market in 2026
max_queries: 5
stream: false
web_search_options:
count: 10
highlight:
enable: true
max_tokens: 300
full_content:
enable: true
max_tokens: 2048
queriesOnly:
summary: Queries Only Mode
value:
model: anthropic/claude-sonnet-4.6
messages:
- role: user
content: What are the latest developments in quantum computing?
mode: queries_only
max_queries: 10
queriesAndSearch:
summary: Queries and Search Mode (no LLM synthesis)
value:
model: anthropic/claude-sonnet-4.6
messages:
- role: user
content: Compare Tesla and BYD electric vehicle sales in 2026
mode: queries_and_search
max_queries: 8
web_search_options:
count: 5
highlight:
enable: true
max_tokens: 300
multiTurn:
summary: Multi-turn Conversation
value:
model: anthropic/claude-sonnet-4.6
messages:
- role: user
content: Who are the major players in the current chip market?
- role: assistant
content: >-
Major players include NVIDIA, AMD, Intel, Broadcom,
Google, Amazon, etc.
- role: user
content: What are their latest developments?
max_queries: 10
stream: false
withDomainFilter:
summary: With Domain and Time Filters
value:
model: anthropic/claude-sonnet-4.6
messages:
- role: user
content: Latest central bank interest rate decisions globally
max_queries: 15
stream: false
web_search_options:
count: 10
include_domains:
- reuters.com
exclude_domains:
- medium.com
include_text:
- interest rate
- central bank
exclude_text:
- opinion
- rumor
time_basis: published
start_time: '2025-01-01T00:00:00Z'
end_time: '2025-01-31T23:59:59Z'
highlight:
enable: true
max_tokens: 300
format: markdown
safesearch: strict
full_content:
enable: true
max_tokens: 2048
responses:
'200':
description: >-
Successful broad search response. When `stream=false`, returns a
single `chat.completion` object with `queries` and `search_results`
at the top level. When `stream=true`, returns a stream of
`chat.completion.chunk` objects with types: `queries` (generated
sub-queries), `search_done` (search results), `content` (incremental
content), `finish` (completion signal), and `usage` (token usage).
content:
application/json:
schema:
oneOf:
- $ref: '#/components/schemas/BroadSearchResponse'
- $ref: '#/components/schemas/BroadSearchChunk'
discriminator:
propertyName: object
mapping:
chat.completion:
$ref: '#/components/schemas/BroadSearchResponse'
chat.completion.chunk:
$ref: '#/components/schemas/BroadSearchChunk'
examples:
nonStreaming:
summary: Non-streaming response (chat.completion)
value:
request_id: 20260403120000001ABCDE12345
object: chat.completion
created: 1775361600
model: anthropic/claude-sonnet-4.6
choices:
- index: 0
finish_reason: stop
message:
role: assistant
content: >-
## Global AI Chip Market in 2026: Latest Trends &
Major Players
**The AI chip market was valued at approximately
$102.9 billion in 2025, and is projected to grow at
a 29.4% CAGR to $1.35 trillion by 2035.** NVIDIA
continues to dominate with over 80% data center
market share, but the custom ASIC chip wave is
accelerating challenges to its position.[^1][^6]
### Major Players Overview
| Company | Core Strength | Key 2026 Developments |
|---|---|---|
| **NVIDIA** | Data center GPU, >80% market share |
Blackwell/Rubin roadmap, $1T revenue target |
| **Broadcom** | Custom ASIC chips | Expected to
capture 60% of custom AI chip market by 2027 |
| **AMD** | GPU alternative | MI450 chip volume ramp
in H2, OpenAI/Meta large orders |
| **Intel** | CPU + foundry | Gaudi 3 launch, Xeon 6
selected for Nvidia DGX systems |
| **Google** | In-house TPU (Ironwood) | Surging HBM
demand |
| **Amazon AWS** | Trainium series | Anthropic
training Claude on over 1M Trainium2 chips |
queries:
- 2026 global AI chip market trends
- AI chip market share 2026
- top AI chip manufacturers 2026
- NVIDIA AI chip dominance 2026
- AMD AI chip strategy 2026
search_results:
- query: 2026 global AI chip market trends
results:
- title: >-
AI Chip Market Size to Exceed USD 1354.35 Billion
by 2035
url: >-
http://globenewswire.com/news-release/2026/04/01/...
highlight: >-
According to the SNS Insider, The AI Chip Market
Size was valued at USD 102.89 Billion in 2025 and
is expected to reach USD 1354.35 Billion by
2035...
full_content: ''
authors: SNS Insider pvt ltd
time_published: '2026-04-01T07:30:00Z'
time_last_crawled: '2026-04-02T03:18:46Z'
latency: 69
- query: AI chip market share 2026
results:
- title: Another result title...
url: https://example.com/...
highlight: ...
full_content: ''
authors: ...
time_published: '2026-03-28T00:00:00Z'
time_last_crawled: '2026-03-29T12:00:00Z'
latency: 72
meta:
usage:
num_search_queries: 20
prompt_tokens: 35351
completion_tokens: 942
total_tokens: 36293
latency: 24048
streamQueries:
summary: 'Streaming chunk — queries (type: queries)'
value:
type: queries
request_id: 20260403120000002XYZAB67890
object: chat.completion.chunk
created: 1775361600
model: anthropic/claude-sonnet-4.6
queries:
- latest quantum computing developments 2026
- quantum computing breakthroughs 2026
streamSearchDone:
summary: 'Streaming chunk — search results (type: search_done)'
value:
type: search_done
request_id: 20260403120000002XYZAB67890
object: chat.completion.chunk
created: 1775361600
model: anthropic/claude-sonnet-4.6
search_results:
- query: latest quantum computing developments 2026
results:
- title: Stocks to Gain From Quantum Computing in 2025
url: https://www.nasdaq.com/articles/...
highlight: >-
This year has seen quantum computing being pushed
from lab interests toward practical deployments...
full_content: ''
authors: ''
time_published: '2026-03-15T00:00:00Z'
time_last_crawled: '2026-03-16T12:00:00Z'
latency: 65
streamContent:
summary: 'Streaming chunk — content (type: content)'
value:
type: content
request_id: 20260403120000002XYZAB67890
object: chat.completion.chunk
created: 1775361600
model: anthropic/claude-sonnet-4.6
choices:
- index: 0
delta:
content: '**Quantum computing'
finish_reason: null
streamFinish:
summary: 'Streaming chunk — finish (type: finish)'
value:
type: finish
request_id: 20260403120000002XYZAB67890
object: chat.completion.chunk
created: 1775361600
model: anthropic/claude-sonnet-4.6
choices:
- index: 0
delta: {}
finish_reason: stop
streamUsage:
summary: 'Streaming chunk — usage (type: usage)'
value:
type: usage
request_id: 20260403120000002XYZAB67890
object: chat.completion.chunk
created: 1775361600
model: anthropic/claude-sonnet-4.6
meta:
usage:
num_search_queries: 2
prompt_tokens: 2664
completion_tokens: 78
total_tokens: 2742
latency: 3565
'400':
description: >-
Missing parameter query — Returned when a required parameter is
missing.
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
code: 400
msg: Missing parameter query
'401':
$ref: '#/components/responses/Unauthorized'
'403':
$ref: '#/components/responses/InsufficientBalance'
'429':
$ref: '#/components/responses/RateLimited'
'500':
$ref: '#/components/responses/InternalError'
security:
- apiKeyAuth: []
- bearerAuth: []
components:
schemas:
BroadSearchRequest:
type: object
required:
- messages
description: Request body for the Broad Search API.
properties:
model:
type: string
enum:
- anthropic/claude-opus-4.8
- anthropic/claude-opus-4.6
- anthropic/claude-sonnet-4.6
- anthropic/claude-haiku-4.5
- google/gemini-3.5-flash
- google/gemini-3.1-pro-preview
- google/gemini-3.1-flash-lite
- google/gemini-3-flash-preview
- openai/gpt-5.5-pro
- openai/gpt-5.5
- openai/gpt-5.4
- moonshotai/kimi-k2.6
- moonshotai/kimi-k2.5
- minimax/minimax-m2.5
- qwen/qwen3.6-plus
default: anthropic/claude-sonnet-4.6
description: The model to use for query decomposition and response synthesis.
messages:
type: array
description: >-
A list of messages comprising the conversation so far. User and
assistant messages in chronological order for multi-turn
conversations.
items:
$ref: '#/components/schemas/ChatMessage'
mode:
type: string
enum:
- queries_only
- queries_and_search
- full
default: full
description: >-
Controls the execution depth. `queries_only`: only decompose the
message into sub-queries without performing searches;
`queries_and_search`: decompose into sub-queries and return search
results without LLM synthesis; `full`: decompose, search, and
synthesize a final response using the LLM.
max_queries:
type: integer
minimum: 1
maximum: 30
default: 30
description: Maximum number of sub-queries to generate.
web_search_options:
allOf:
- $ref: '#/components/schemas/ChatWebSearchOptions'
description: >-
Search-related options. Shares the same parameters and defaults as
the Search API, except `highlight.max_tokens` defaults to 256
(instead of 512). Queries are automatically generated from the
messages.
stream:
type: boolean
default: false
description: >-
Whether to enable streaming output. When `true`, returns
`chat.completion.chunk` objects incrementally with types: `queries`,
`search_done`, `content`, `finish`, and `usage`.
BroadSearchResponse:
type: object
description: A non-streaming Broad Search response. Returned when `stream=false`.
required:
- request_id
- object
- created
- model
- choices
properties:
request_id:
type: string
description: The unique identifier for this request.
object:
type: string
enum:
- chat.completion
description: >-
The object type, always `chat.completion` for non-streaming
responses.
created:
type: number
description: Unix timestamp (in seconds) of when the completion was created.
model:
type: string
description: The model used for this completion.
choices:
type: array
description: A list of completion choices containing the synthesized response.
items:
$ref: '#/components/schemas/ChatCompletionChoice'
queries:
type: array
items:
type: string
description: >-
The list of sub-queries automatically generated from the user
message by the system.
search_results:
type: array
description: >-
Search results grouped by query. Each auto-generated sub-query has a
corresponding result group.
items:
$ref: '#/components/schemas/BroadSearchResultGroup'
meta:
$ref: '#/components/schemas/BroadSearchMeta'
warning:
type: string
nullable: true
description: Warning message, if any.
BroadSearchChunk:
type: object
description: >-
A streaming chunk of a Broad Search response. Returned when
`stream=true`. The `type` field indicates the chunk kind: `queries`
(auto-generated sub-queries), `search_done` (search results available),
`content` (incremental content), `finish` (generation complete), `usage`
(token usage summary). The `queries` type is unique to Broad Search and
is not present in the Web Chat API.
required:
- request_id
- object
- created
- model
properties:
type:
type: string
enum:
- queries
- search_done
- content
- finish
- usage
description: >-
The type of this streaming chunk. `queries` is unique to Broad
Search and contains the auto-generated sub-queries.
request_id:
type: string
description: The unique identifier for this request.
object:
type: string
enum:
- chat.completion.chunk
description: >-
The object type, always `chat.completion.chunk` for streaming
responses.
created:
type: number
description: Unix timestamp (in seconds) of when the chunk was created.
model:
type: string
description: The model used for this completion.
queries:
type: array
items:
type: string
description: Auto-generated sub-queries. Present only in `queries` type chunks.
choices:
type: array
description: Incremental choices. Present in `content` and `finish` chunks.
items:
$ref: '#/components/schemas/ChatCompletionChunkChoice'
search_results:
type: array
description: Search results grouped by query. Present in `search_done` chunks.
items:
$ref: '#/components/schemas/BroadSearchResultGroup'
meta:
$ref: '#/components/schemas/BroadSearchMeta'
ErrorResponse:
type: object
properties:
code:
type: integer
description: Business status code. Non-zero values indicate an error.
msg:
type: string
description: A human-readable message describing the error.
required:
- code
- msg
ChatMessage:
description: A single message in the conversation. Discriminated by `role`.
oneOf:
- $ref: '#/components/schemas/SystemMessage'
- $ref: '#/components/schemas/UserMessage'
- $ref: '#/components/schemas/AssistantMessage'
- $ref: '#/components/schemas/ToolMessage'
discriminator:
propertyName: role
mapping:
system:
$ref: '#/components/schemas/SystemMessage'
user:
$ref: '#/components/schemas/UserMessage'
assistant:
$ref: '#/components/schemas/AssistantMessage'
tool:
$ref: '#/components/schemas/ToolMessage'
ChatWebSearchOptions:
type: object
description: >-
Search-related options for the Web Chat API. All parameters are optional
and share the same semantics and defaults as the Search API. The query
is automatically generated from the messages and does not need to be
provided.
properties:
count:
type: integer
minimum: 1
maximum: 100
description: Number of search results to return.
include_domains:
type: array
items:
type: string
description: Domains to include in search results.
exclude_domains:
type: array
items:
type: string
description: Domains to exclude from search results.
include_text:
type: array
items:
type: string
maxItems: 5
description: Strings that must appear in the result page text.
exclude_text:
type: array
items:
type: string
maxItems: 5
description: Strings that must not appear in the result page text.
time_basis:
type: string
enum:
- auto
- published
- crawled
description: Determines which time field is used for time filtering.
start_time:
type: string
format: date-time
description: Start time for filtering results. ISO 8601 format.
end_time:
type: string
format: date-time
description: End time for filtering results. ISO 8601 format.
highlight:
$ref: '#/components/schemas/HighlightOptions'
format:
type: string
enum:
- markdown
- text
description: Controls the formatting of highlight outputs.
safesearch:
type: string
enum:
- 'off'
- strict
description: Controls filtering of explicit/adult content.
full_content:
$ref: '#/components/schemas/FullContentOptions'
ChatCompletionChoice:
type: object
description: A single completion choice in a non-streaming response.
properties:
index:
type: number
description: The index of this choice in the list. Usually 0.
finish_reason:
type: string
enum:
- stop
- length
- tool_calls
- content_filter
- error
description: >-
The reason the model stopped generating. `stop` indicates normal
completion.
message:
$ref: '#/components/schemas/ChatCompletionMessage'
BroadSearchResultGroup:
type: object
description: >-
A group of search results for a single auto-generated sub-query in Broad
Search.
properties:
query:
type: string
description: The auto-generated sub-query that produced these results.
results:
type: array
description: The search results for this sub-query.
items:
$ref: '#/components/schemas/BroadSearchResultItem'
latency:
type: integer
description: Search latency for this query in milliseconds.
BroadSearchMeta:
type: object
description: Metadata for the Broad Search response.
properties:
usage:
$ref: '#/components/schemas/BroadSearchUsage'
latency:
type: integer
description: Total request latency in milliseconds.
ChatCompletionChunkChoice:
type: object
description: A single choice in a streaming chunk.
properties:
index:
type: number
description: The index of this choice. Usually 0.
delta:
$ref: '#/components/schemas/ChatCompletionDelta'
finish_reason:
type: string
nullable: true
enum:
- stop
- length
- tool_calls
- content_filter
- error
- null
description: >-
The reason generation stopped. `null` for intermediate chunks; set
in the `finish` chunk.
SystemMessage:
type: object
required:
- role
- content
description: A system prompt message that sets the behavior or context for the model.
properties:
role:
type: string
enum:
- system
description: The role of the message author. Always `system`.
content:
type: string
description: The system prompt content.
UserMessage:
type: object
required:
- role
- content
description: A message from the user.
properties:
role:
type: string
enum:
- user
description: The role of the message author. Always `user`.
content:
description: >-
The content of the message. Can be a plain string or an array of
content blocks (text and/or image).
oneOf:
- type: string
- type: array
items:
$ref: '#/components/schemas/ChatContentBlock'
AssistantMessage:
type: object
required:
- role
description: >-
A message from the assistant. May contain text content, tool calls, or
both. When replaying a multi-turn conversation with tool use, include
the assistant's `tool_calls` so the model can match them with the
subsequent `tool` messages.
properties:
role:
type: string
enum:
- assistant
description: The role of the message author. Always `assistant`.
content:
type: string
nullable: true
description: >-
The assistant's text content. May be `null` or omitted when the
assistant only produces tool calls.
tool_calls:
type: array
description: >-
Tool calls generated by the model. Each tool call must be answered
by a corresponding `tool` message with a matching `tool_call_id`.
items:
$ref: '#/components/schemas/ChatToolCall'
ToolMessage:
type: object
required:
- role
- tool_call_id
- content
description: >-
A tool result message, providing the output of a tool call back to the
model. The `tool_call_id` must match the `id` of a preceding
`tool_calls` entry in an assistant message.
properties:
role:
type: string
enum:
- tool
description: The role of the message author. Always `tool`.
tool_call_id:
type: string
description: >-
The ID of the tool call this message is responding to. Must match an
`id` from a preceding assistant message's `tool_calls`.
content:
type: string
description: The tool output, typically a JSON string with the function result.
HighlightOptions:
type: object
description: Controls highlight extraction from result pages.
properties:
enable:
type: boolean
default: true
description: If true, returns query-relevant highlight in each result.
max_tokens:
type: integer
default: 512
minimum: 100
maximum: 20000
description: Max tokens returned per highlight.
FullContentOptions:
type: object
description: Controls whether to return the full raw content of each result page.
properties:
enable:
type: boolean
default: false
description: If true, returns full_content for each result.
max_tokens:
type: integer
default: 2048
minimum: 100
maximum: 100000
description: Maximum tokens of full content included per result.
ChatCompletionMessage:
type: object
description: >-
The assistant's response message (non-streaming). For reasoning models,
the thinking process appears as `...` tags within the
`content` field. In streaming mode, reasoning content is delivered via
`delta.reasoning_content` instead.
properties:
role:
type: string
enum:
- assistant
- tool
description: The role of the message author. Usually `assistant`.
content:
type: string
nullable: true
description: >-
The assistant's text content. May be `null` when the model only
produces tool calls, or in refusal scenarios. For reasoning models,
this field may contain `...` tags wrapping the
model's reasoning process before the final answer.
tool_calls:
type: array
description: >-
Tool calls generated by the model. Only present when the model
decides to invoke one or more tools. Each call includes an `id` that
must be referenced by a subsequent `tool` message.
items:
$ref: '#/components/schemas/ChatToolCall'
refusal:
type: string
description: >-
A model-generated refusal message. Only present when the model
refuses a request.
BroadSearchResultItem:
type: object
description: A single search result item within a Broad Search result group.
properties:
title:
type: string
description: The title of the result page.
url:
type: string
description: The URL of the result page.
highlight:
type: string
description: >-
Query-relevant highlight content. Only returned when
`highlight.enable` is `true`.
full_content:
type: string
description: >-
Full page content. Only returned when `full_content.enable` is
`true`.
authors:
type: string
description: Website name or author.
time_published:
type: string
format: date-time
description: Publish time in ISO 8601.
time_last_crawled:
type: string
format: date-time
description: Last crawl time in ISO 8601.
BroadSearchUsage:
type: object
description: >-
Token usage information for the Broad Search response. When
`stream=true`, this is only present in the `usage` type chunk.
required:
- num_search_queries
- prompt_tokens
- completion_tokens
- total_tokens
properties:
num_search_queries:
type: integer
description: Total number of search queries executed across all sub-queries.
prompt_tokens:
type: integer
description: Number of input tokens.
completion_tokens:
type: integer
description: Number of output tokens.
total_tokens:
type: integer
description: Total tokens used (prompt_tokens + completion_tokens).
ChatCompletionDelta:
type: object
description: >-
Incremental content in a streaming chunk. Contains only the new fields
for this chunk.
properties:
role:
type: string
enum:
- assistant
description: The role. Typically included in the first chunk.
reasoning_content:
type: string
description: >-
Incremental reasoning content from reasoning models. Present during
the thinking phase before the final answer. Concatenate across
chunks to build the complete reasoning. Only appears for
reasoning-capable models (`google/gemini-3.1-pro-preview`,
`moonshotai/kimi-k2.6`, `moonshotai/kimi-k2.5`,
`minimax/minimax-m2.5`).
content:
type: string
description: Incremental text content.
tool_calls:
type: array
description: >-
Incremental tool call data. Present when the model invokes tools
during streaming. Each chunk may contain partial function name or
arguments that should be concatenated.
items:
$ref: '#/components/schemas/ChatToolCallDelta'
ChatContentBlock:
type: object
description: A content block within a message. Supports text and image types.
properties:
type:
type: string
enum:
- text
- image_url
description: The type of content block.
text:
type: string
description: The text content. Required when `type` is `text`.
image_url:
type: object
description: Image URL object. Required when `type` is `image_url`.
properties:
url:
type: string
description: The URL of the image.
required:
- url
required:
- type
ChatToolCall:
type: object
required:
- id
- type
- function
description: A tool call generated by the model.
properties:
index:
type: number
description: The index of this tool call in the tool_calls array.
id:
type: string
description: >-
A unique identifier for this tool call. Referenced by the
corresponding `tool` message's `tool_call_id`.
type:
type: string
enum:
- function
description: The type of tool call. Currently only `function` is supported.
function:
$ref: '#/components/schemas/ChatToolCallFunction'
ChatToolCallDelta:
type: object
description: >-
Incremental tool call data in a streaming chunk. The first chunk for a
tool call includes `id`, `type`, and `function.name`. Subsequent chunks
append to `function.arguments`.
properties:
index:
type: number
description: The index of this tool call in the tool_calls array.
id:
type: string
description: >-
The tool call ID. Only present in the first chunk for this tool
call.
type:
type: string
enum:
- function
description: The type of tool call.
function:
type: object
description: Incremental function call data.
properties:
name:
type: string
description: >-
The function name. Only present in the first chunk for this tool
call.
arguments:
type: string
description: >-
Incremental JSON string of the function arguments. Concatenate
across chunks to build the complete arguments.
ChatToolCallFunction:
type: object
required:
- name
- arguments
description: The function invocation details within a tool call.
properties:
name:
type: string
description: The name of the function to call.
arguments:
type: string
description: >-
The arguments to the function, as a JSON string generated by the
model.
responses:
Unauthorized:
description: Invalid API Key — Returned when the API key is missing or invalid.
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
code: 401
msg: Invalid API Key
InsufficientBalance:
description: >-
Insufficient balance in account — Returned when the account balance is
insufficient to complete the request.
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
code: 403
msg: Insufficient balance in account
RateLimited:
description: >-
Exceeding the rate limit — Returned when the request exceeds the
configured rate limit.
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
code: 429
msg: Exceeding the rate limit
InternalError:
description: Internal error — Returned when an unexpected server-side error occurs.
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
code: 500
msg: Internal error
securitySchemes:
apiKeyAuth:
type: apiKey
in: header
name: x-api-key
description: >-
API key used for request authentication. Obtain an API key before using
the API. Note: A payment method is required to use the API.
bearerAuth:
type: http
scheme: bearer
description: >-
Bearer token authentication. Compatible with OpenAI protocol. Pass the
API key as `Authorization: Bearer `.
````