[KongchangAI]
Concept

Function Calling

Core Facts

Timeline (last 90 days)

Nov 18

Products like Beat are most likely built on top of LLMs, connecting to external services through function calling or similar mechanisms.

Unverified50%
Nov 18

Function calling is a core mechanism for large language models to interact with external tools: developers predefine functions and describe their purpose and parameters, and the LLM decides whether to call them and generates the call parameters.

Unverified50%
Nov 17

Large models themselves do not have built-in real-time knowledge of IP reputation; the point of tool calling is to bring external deterministic data sources into the reasoning chain, with the model deciding whether to query and the external service providing the query result.

Unverified50%
Nov 17

In this workflow, the AI Agent uses the Gemini model and calls the threat_intel_lookup tool via Function Calling; the tool input is the source IP, and it returns a fixed JSON structure containing four fields: severity, verdict, suggested action, and a one-sentence summary.

Unverified50%
Nov 16

Many model vendors implement information extraction (structured output) and tool calling within the same API.

Unverified50%
Nov 13

The core of Spacefast's agent integration is a set of AI-oriented tool invocation interfaces (Function Calling / MCP-type protocols), through which agents perform actions such as creating projects, writing files, and executing deployments via structured calls.

Unverified50%
Nov 12

Function Calling refers to a large model outputting a structured tool invocation request (tool name plus JSON parameters), which the application side parses, executes, and returns the result of; it is essentially an output convention on the model side, and the parameter formats of different vendors are not entirely consistent.

Unverified50%
Nov 9

With the function-calling capabilities of large models, developers still need to complete OAuth integration, but can devote more effort to tool descriptions and prompt design, reducing development costs at the intent-parsing layer.

Unverified50%
Nov 5

Compared with early ReAct prompt-based approaches, native function calling is specifically aligned by the model during training, yielding higher accuracy in both parameter construction and tool selection.

Unverified50%
Nov 3

Through mechanisms such as Function Calling or MCP (Model Context Protocol), the Agent can proactively initiate HTTP requests, execute system commands, and read and write files.

Unverified50%

40 more timeline events

All Facts (20)

Verified

Mainstream Agent architectures typically use a large language model as the brain, combined with memory modules, planning modules, and tool interfaces to accomplish complex tasks

90%
Verified

In the tool-calling mechanism, developers attach a tool description list in JSON Schema format when calling the LLM API, and the model outputs call parameters in structured JSON format

90%
Verified

Function Calling (Tool Use) is a standardized mechanism for large models to invoke external tools, where developers pre-declare tool names, descriptions, and parameter schemas, and the model autonomously determines when to make a call.

90%
Verified

The underlying mechanism of Tool Use is that developers pre-define the name, description, and parameter structure (Schema) of tools, and the model determines during inference when and which tool to call and outputs structured invocation instructions.

90%
Verified

The tool-calling (Function Calling) mechanism allows the model to invoke external tools during reasoning, such as file read/write operations, terminal commands, and search APIs.

90%
Verified

OpenAI first introduced the Function Calling feature in June 2023, after which Anthropic's Claude and Google's Gemini also successively added support.

90%
Verified

OpenAI's Function Calling and Anthropic's Tool Use API are both standardized implementations of the ReAct pattern.

90%
Verified

OpenAI launched the Function Calling specification in 2023, allowing developers to describe function interfaces in JSON Schema format, with the model autonomously deciding when to call them and what parameters to pass.

90%
Verified

Each tool in MCP has a strict JSON Schema description, including the name, functional description, and parameter specifications

90%
Verified

Agent planning capabilities are typically implemented through the ReAct (Reasoning + Acting) framework.

90%
Verified

A typical Agent is composed of a cycle of 'Perception—Planning—Action—Reflection'.

85%
Verified

Before MCP appeared, if developers wanted AI to call external services, they usually had to write Function Calling code separately for each service, and the interfaces of different model providers were mutually incompatible.

80%
Verified

The core mechanism of Function Calling is that developers describe external function signatures in JSON Schema format, and the model outputs structured JSON invocation instructions that conform to the schema specification.

80%
Verified

The underlying operational logic of an Agent Loop is an iterative cycle of 'perception–planning–action–observation.'

80%
Verified

The Function Calling mechanism allows LLMs to declare in structured JSON format the external tools and parameters they need to invoke, and the tool's return value is injected as-is into the next round of model input.

80%
Verified

DeepSeek R1 does not support Function Calling by default, whereas DeepSeek V3 supports Function Calling.

80%
Verified

Function Calling is the core mechanism for large models to interact with external systems, allowing the model to call predefined functions after identifying user intent.

80%
Verified

在MCP出现之前,AI调用外部工具主要依赖OpenAI的Function Calling机制或各框架自定义的插件系统

80%
Verified

Tool-Augmented AI is an important development direction for current large model applications, with mechanisms such as OpenAI's Function Calling and Anthropic's Tool Use underpinning this paradigm at the foundational level.

80%
Verified

The principle of Function Calling is that, in addition to generating natural language text during inference, the model can also output structured function-call instructions (JSON containing the function name and parameters).

80%

Source Articles