Fill-in-the-middle code completion endpoints.
openapi: 3.1.0
info:
title: Inception Chat FIM API
version: 1.0.0
description: Inception Labs LLM API — chat, FIM, and edit completions powered by Mercury diffusion language models. OpenAI-compatible request/response shapes with extensions for diffusion-specific features.
contact:
name: Inception Labs
url: https://docs.inceptionlabs.ai
email: support@inceptionlabs.ai
license:
name: Proprietary
x-logo:
url: https://docs.inceptionlabs.ai/logo.png
servers:
- url: https://api.inceptionlabs.ai
description: Production
security:
- BearerAuth: []
tags:
- name: FIM
description: Fill-in-the-middle code completion endpoints.
paths:
/v1/fim/completions:
post:
summary: Create a fill-in-the-middle completion
description: Generate a code completion given a `prompt` (prefix) and optional `suffix`. Designed for IDE-style inline completion. Returns a `FimCompletion` object, or a server-sent events stream of `FimCompletionChunk` deltas when `stream=true`. Tool calling and function calling are not supported.
operationId: createFimCompletion
requestBody:
content:
application/json:
schema:
$ref: '#/components/schemas/FIMCompletionRequest'
example:
model: mercury-edit-2
prompt: "def fibonacci(n: int) -> int:\n if n <= 1:\n return n\n return "
suffix: '
print(fibonacci(10))
'
max_tokens: 256
required: true
responses:
'200':
description: 'Successful response. Returns a JSON object when `stream=false`, or a server-sent events stream of `TextCompletionChunk` objects (terminated by `data: [DONE]`) when `stream=true`.'
content:
application/json:
schema:
$ref: '#/components/schemas/FIMCompletionResponse'
example:
id: cmpl-7a2b3c4d5e
object: text_completion
created: 1745798400
model: mercury-edit-2
choices:
- index: 0
finish_reason: stop
text: fibonacci(n - 1) + fibonacci(n - 2)
usage:
prompt_tokens: 24
completion_tokens: 14
total_tokens: 38
reasoning_tokens: 0
cached_input_tokens: 0
text/event-stream:
schema:
$ref: '#/components/schemas/TextCompletionChunk'
x-stainless-stream-type: sse
x-stainless-stream-event-schema:
$ref: '#/components/schemas/TextCompletionChunk'
x-stainless-stream-terminator: '[DONE]'
'400':
description: Bad Request — invalid parameters.
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
message: You exceeded the maximum context length for this model of 128000. Please reduce the length of the messages or completion.
type: invalid_request_error
param: messages
code: context_length_exceeded
'401':
description: Unauthorized — missing or invalid API key.
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
message: Incorrect API key provided
type: authentication_error
param: null
code: invalid_api_key
'402':
description: Payment Required — billing inactive or quota exceeded.
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
message: Account is inactive
type: account_error
param: null
code: account_error
'404':
description: Not Found — model not available for this account.
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
message: model `jupyter-2` not found
type: invalid_request_error
param: model
code: model_not_found
'429':
description: Too Many Requests — rate limit exceeded.
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
message: Rate limit exceeded. Please try again later.
type: rate_limit_error
param: null
code: rate_limit_reached
'500':
description: Internal Server Error.
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorResponse'
example:
error:
message: The server had an error while processing your request.
type: server_error
param: null
code: server_error
security:
- BearerAuth: []
tags:
- FIM
x-codeSamples:
- lang: python
label: Python
source: "import os\nfrom inceptionai import Inception\n\nclient = Inception(\n api_key=os.environ.get(\"INCEPTION_API_KEY\"), # defaults to this env var; can be omitted\n)\n\nfim_completion = client.fim.completions.create(\n model=\"mercury-edit-2\",\n prompt=\"def fibonacci(n: int) -> int:\\n if n <= 1:\\n return n\\n return \",\n suffix=\"\\n\\nprint(fibonacci(10))\\n\",\n max_tokens=256,\n)\nprint(fim_completion)"
- lang: typescript
label: TypeScript
source: "import Inception from 'inceptionai';\n\nconst client = new Inception({\n apiKey: process.env['INCEPTION_API_KEY'], // defaults to this env var; can be omitted\n});\n\nconst fimCompletion = await client.fim.completions.create({\n model: 'mercury-edit-2',\n prompt: 'def fibonacci(n: int) -> int:\\n if n <= 1:\\n return n\\n return ',\n suffix: '\\n\\nprint(fibonacci(10))\\n',\n max_tokens: 256\n});\nconsole.log(fimCompletion);"
components:
schemas:
FIMUsage:
properties:
prompt_tokens:
type: integer
title: Prompt Tokens
reasoning_tokens:
type: integer
title: Reasoning Tokens
completion_tokens:
type: integer
title: Completion Tokens
total_tokens:
type: integer
title: Total Tokens
cached_input_tokens:
type: integer
title: Cached Input Tokens
type: object
required:
- prompt_tokens
- reasoning_tokens
- completion_tokens
- total_tokens
- cached_input_tokens
title: FIMUsage
description: Usage for FIM completions.
TextChunkChoice:
description: Represents a choice in a streaming text completion response chunk.
properties:
index:
title: Index
type: integer
text:
anyOf:
- type: string
- type: 'null'
default: null
description: The text in this chunk.
title: Text
finish_reason:
type:
- string
- 'null'
enum:
- stop
- length
- content_filter
- null
description: The reason the model stopped generating tokens.
required:
- index
title: TextChunkChoice
type: object
FIMCompletionRequest:
properties:
prompt:
type: string
title: Prompt
description: The prompt to complete.
suffix:
type: string
title: Suffix
description: The suffix to complete.
model:
type: string
title: Model
description: The model to use for the FIM completion.
max_tokens:
type: integer
title: Max Tokens
description: Maximum number of tokens to generate.
default: 512
minimum: 1
maximum: 8192
top_p:
type: number
title: Top P
description: Float that controls the cumulative probability of the top tokens to consider.
default: 1.0
minimum: 0.0
maximum: 1.0
top_k:
type: integer
title: Top K
description: Limits sampling to the `k` most likely tokens. Must be `-1` (disables the cutoff and considers all tokens) or an integer from 1 to 1000; other values such as 0 are rejected.
default: -1
minimum: -1
maximum: 1000
frequency_penalty:
type: number
title: Frequency Penalty
description: Number between -2 and 2. Positive values penalize tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim.
default: 0.0
minimum: -2.0
maximum: 2.0
presence_penalty:
type: number
title: Presence Penalty
description: Number between -2 and 2. Positive values penalize tokens based on whether they have appeared in the text so far, increasing the model's likelihood to talk about new topics.
default: 1.5
minimum: -2.0
maximum: 2.0
repetition_penalty:
type: number
title: Repetition Penalty
description: Penalizes tokens that have already appeared in the generated text. Must be greater than 0. Values greater than 1.0 discourage repetition; 1.0 applies no penalty.
default: 1.0
minimum: 0
stop:
items:
type: string
type: array
title: Stop
description: A list of sequences where the API will stop generating further tokens. The returned text will not contain the stop sequences. Defaults to common code-block boundaries.
default:
- '
'
- "\n \n"
stream:
type: boolean
title: Stream
description: Whether to stream the response.
default: false
stream_options:
$ref: '#/components/schemas/StreamOptions'
description: Options that control streaming behavior.
type: object
required:
- prompt
- model
title: FIMCompletionRequest
ErrorResponse:
type: object
title: ErrorResponse
required:
- error
properties:
error:
$ref: '#/components/schemas/ErrorObject'
example:
error:
message: You exceeded the maximum context length for this model of 128000. Please reduce the length of the messages or completion.
type: invalid_request_error
param: messages
code: context_length_exceeded
FIMCompletionResponse:
properties:
id:
type: string
title: Id
object:
type: string
const: text_completion
title: Object
default: text_completion
created:
type: integer
title: Created
model:
type: string
title: Model
choices:
items:
$ref: '#/components/schemas/TextCompletionChoice'
type: array
title: Choices
usage:
$ref: '#/components/schemas/FIMUsage'
warning:
anyOf:
- type: string
- type: 'null'
title: Warning
type: object
required:
- id
- created
- model
- choices
- usage
- object
title: FIMCompletionResponse
example:
id: cmpl-7a2b3c4d5e
object: text_completion
created: 1745798400
model: mercury-edit-2
choices:
- index: 0
finish_reason: stop
text: fibonacci(n - 1) + fibonacci(n - 2)
usage:
prompt_tokens: 24
completion_tokens: 14
total_tokens: 38
reasoning_tokens: 0
cached_input_tokens: 0
TextCompletionChoice:
properties:
index:
type: integer
title: Index
text:
type: string
title: Text
finish_reason:
type:
- string
- 'null'
enum:
- stop
- length
- content_filter
- null
description: The reason the model stopped generating tokens.
type: object
required:
- index
- text
- finish_reason
title: TextCompletionChoice
description: Choice for FIM completions.
TextCompletionChunk:
description: Model for a single chunk in a streaming text completion response.
properties:
id:
title: Id
type: string
object:
const: text_completion
default: text_completion
title: Object
type: string
created:
title: Created
type: integer
model:
title: Model
type: string
choices:
items:
$ref: '#/components/schemas/TextChunkChoice'
title: Choices
type: array
required:
- id
- created
- model
- choices
- object
title: TextCompletionChunk
type: object
StreamOptions:
properties:
include_usage:
type: boolean
title: Include Usage
description: 'If true, an additional chunk is streamed before the `data: [DONE]` message containing the token usage statistics for the entire request. Inside that chunk, `choices` is an empty array and the `usage` field is populated.'
default: false
type: object
title: StreamOptions
description: Options for controlling streaming behavior. Only used when `stream=true`.
required: []
ErrorObject:
type: object
title: ErrorObject
required:
- message
- type
properties:
message:
type: string
description: Human-readable error message.
type:
type: string
description: Error category, e.g. `invalid_request_error`, `rate_limit_error`.
param:
type:
- string
- 'null'
description: Offending parameter, if applicable.
code:
type:
- string
- 'null'
description: Machine-readable error code.
securitySchemes:
BearerAuth:
type: http
scheme: bearer
description: 'API key provided as a Bearer token: `Authorization: Bearer <api_key>`. Get an API key at https://platform.inceptionlabs.ai.'