All types are TypeScript interfaces exported from infinium-o2.

Core Types

TaskData

The central data structure sent to the Infinium API. Represents a complete trace of an agent’s work.

import type { TaskData } from 'infinium-o2';

Field	Type	Required	Description
name	string	Yes	Short task name (max 500 chars)
description	string	Yes	What the agent did (max 10,000 chars)
currentDatetime	string	Yes	ISO 8601 timestamp
duration	number	Yes	Wall-clock time in seconds (0-86,400)
timeTracking	TimeTracking	No	Start/end timestamps
customer	Customer	No	Customer context
support	Support	No	Support ticket context
sales	Sales	No	Sales pipeline context
marketing	Marketing	No	Marketing campaign context
content	Content	No	Content creation context
research	Research	No	Research task context
project	Project	No	Project management context
development	Development	No	Software development context
executive	Executive	No	Executive/meeting context
general	General	No	General-purpose context
llmUsage	LlmUsage	No	Aggregate token/cost stats
inputSummary	string	No	Summary of agent input
outputSummary	string	No	Summary of agent output
outcome	string	No	Outcome description
steps	ExecutionStep[]	No	Ordered execution steps
expectedOutcome	ExpectedOutcome	No	Maestro’s grading rubric
environment	EnvironmentContext	No	Runtime metadata
errors	ErrorDetail[]	No	Errors encountered

ApiResponse<T>

Returned by sendTask(), sendTaskData(), and getInterpretedTaskResult().

import type { ApiResponse } from 'infinium-o2';

Field	Type	Description
success	boolean	Whether the request succeeded
statusCode	number | undefined	HTTP status code
message	string	Response message
data	T | undefined	Response payload (includes traceId on success)

BatchResult

Returned by sendBatch() and sendBatchSequential().

import type { BatchResult } from 'infinium-o2';

Field	Type	Description
totalTasks	number	Total tasks in the batch
successfulTasks	number	Number sent successfully
failedTasks	number	Number that failed
results	ApiResponse[]	Individual response per task
errors	string[]	Error messages for failures

PromptContent

Returned by getPrompt().

import type { PromptContent } from 'infinium-o2';

Field	Type	Description
promptId	string	The prompt’s UUID
name	string	Display name
version	number	Version number
content	string	Raw template content
createdAt	string	ISO timestamp of version creation
renderedContent	string | undefined	Content with variables substituted

---

Trace Enrichment Types

ExecutionStep

Represents one thing the agent did.

import type { ExecutionStep } from 'infinium-o2';

Field	Type	Required	Description
stepNumber	number	Yes	Sequential step number
action	string	Yes	Action type (e.g., 'llm_inference', 'tool_use', 'decision')
description	string	Yes	What this step did
durationMs	number	No	Step duration in milliseconds
inputPreview	string	No	Preview of step input (max 500 chars)
outputPreview	string	No	Preview of step output (max 500 chars)
error	ErrorDetail	No	Error that occurred
toolCall	ToolCall	No	Tool/API invocation details
llmCall	LlmCall	No	LLM invocation details
metadata	Record<string, unknown>	No	Arbitrary metadata

ErrorDetail

A factual error record.

import type { ErrorDetail } from 'infinium-o2';

Field	Type	Required	Description
errorType	string	Yes	Error class name
message	string	Yes	Error message
recoverable	boolean	No	Whether the agent recovered (default: false)
retryCount	number	No	Number of retries attempted (default: 0)
stackTrace	string	No	Full stack trace
errorCode	string	No	Application error code

ToolCall

A tool or API invocation.

import type { ToolCall } from 'infinium-o2';

Field	Type	Required	Description
toolName	string	Yes	Name of the tool
durationMs	number	No	Call duration in milliseconds
inputSummary	string	No	Summary of input
outputSummary	string	No	Summary of output
error	ErrorDetail	No	Error details if the call failed
httpStatus	number	No	HTTP status code

LlmCall

A single LLM invocation.

import type { LlmCall } from 'infinium-o2';

Field	Type	Required	Description
model	string	Yes	Model identifier
provider	string	No	Provider name (openai, anthropic, google, xai)
promptTokens	number	No	Input token count
completionTokens	number	No	Output token count
latencyMs	number	No	Call latency in milliseconds
temperature	number	No	Temperature parameter
purpose	string	No	What this LLM call was for

LlmUsage

Aggregate token and cost statistics across the entire trace.

import type { LlmUsage } from 'infinium-o2';

Field	Type	Description
model	string | undefined	Primary model used
provider	string | undefined	Primary provider
promptTokens	number | undefined	Total input tokens
completionTokens	number | undefined	Total output tokens
totalTokens	number | undefined	Total tokens (prompt + completion)
estimatedCostUsd	number | undefined	Estimated cost in USD
apiCallsCount	number | undefined	Number of API calls made
totalLatencyMs	number | undefined	Sum of all call latencies
calls	LlmCall[] | undefined	Individual call details

ExpectedOutcome

Defines what Maestro should evaluate the trace against.

import type { ExpectedOutcome } from 'infinium-o2';

Field	Type	Required	Description
taskObjective	string	Yes	The goal of the task
requiredDeliverables	string[]	No	What the agent should produce
constraints	string[]	No	Rules the agent should follow
acceptanceCriteria	string[]	No	How to judge success

EnvironmentContext

Runtime metadata about the execution environment.

import type { EnvironmentContext } from 'infinium-o2';

Field	Type	Description
framework	string | undefined	Framework name (e.g., 'langchain', 'crewai')
frameworkVersion	string | undefined	Framework version
nodeVersion	string | undefined	Node.js version
sdkVersion	string | undefined	Infinium SDK version (auto-populated)
runtime	string | undefined	Runtime environment
region	string | undefined	Deployment region
customTags	Record<string, string> | undefined	Custom key-value tags

---

Configuration Types

ClientConfig

import type { ClientConfig } from 'infinium-o2';

Field	Type	Required	Description
agentId	string	Yes	Agent UUID
agentSecret	string	Yes	Agent secret key
baseUrl	string	No	API base URL
timeout	number	No	Request timeout in ms
maxRetries	number	No	Max retry attempts
enableRateLimiting	boolean	No	Enable rate limiting
requestsPerSecond	number	No	Rate limit threshold
userAgent	string	No	Custom User-Agent
enableLogging	boolean	No	Enable logging
logLevel	string	No	Log level

RequestOptions

import type { RequestOptions } from 'infinium-o2';

Field	Type	Description
timeout	number | undefined	Override timeout for this request
retries	number | undefined	Override retries for this request
retryDelay	number | undefined	Override retry delay
headers	Record<string, string> | undefined	Additional headers

RetryConfig

import type { RetryConfig } from 'infinium-o2';

Field	Type	Description
attempts	number	Max retry attempts
delay	number	Base delay in ms
factor	number | undefined	Backoff multiplier (default: 2)
maxDelay	number | undefined	Max delay cap in ms (default: 30000)
retryCondition	function | undefined	Custom retry predicate

RateLimiterConfig

import type { RateLimiterConfig } from 'infinium-o2';

Field	Type	Description
requestsPerSecond	number	Requests per second
burstSize	number | undefined	Max burst (default: 2x requestsPerSecond)