Documentation / @agentick/core / ModelMetadata

Interface: ModelMetadata

Defined in: core/src/model/model.ts:239

Properties

capabilities

capabilities: ModelCapabilities[]

Defined in: core/src/model/model.ts:246

---

contextWindow?

optional contextWindow: number

Defined in: core/src/model/model.ts:249

Context window size in tokens (adapter-provided takes precedence over catalog)

---

description?

optional description: string

Defined in: core/src/model/model.ts:242

---

id

id: string

Defined in: core/src/model/model.ts:240

---

isReasoningModel?

optional isReasoningModel: boolean

Defined in: core/src/model/model.ts:259

Whether this is a reasoning model (extended thinking)

---

maxOutputTokens?

optional maxOutputTokens: number

Defined in: core/src/model/model.ts:251

Maximum output tokens per response

---

model?

optional model: string

Defined in: core/src/model/model.ts:241

---

provider?

optional provider: string

Defined in: core/src/model/model.ts:244

---

supportsStructuredOutput?

optional supportsStructuredOutput: boolean

Defined in: core/src/model/model.ts:257

Whether the model supports structured output via JSON schema

---

supportsToolUse?

optional supportsToolUse: boolean

Defined in: core/src/model/model.ts:255

Whether the model supports tool/function calling

---

supportsVision?

optional supportsVision: boolean

Defined in: core/src/model/model.ts:253

Whether the model supports vision/image input

---

tokenEstimator?

optional tokenEstimator: TokenEstimator

Defined in: core/src/model/model.ts:262

Token estimator function (e.g., tiktoken). If provided, used instead of default char/4 heuristic.

---

type?

optional type: "image" | "language" | "embedding" | "vision"

Defined in: core/src/model/model.ts:245

---

version?

optional version: string

Defined in: core/src/model/model.ts:243