# Inner Loop Agents

DevFeed: [Inner Loop Agents](<https://devfeed.tech/articles/inner-loop-agents-33461.md>)

Original publisher: [Read original article](<https://timkellogg.me/blog/2025/04/19/inner-loops>)

Published: 2025-04-19T00:00:00Z

Content type: opinion

Language: en

Sources: [Tim Kellogg](<https://devfeed.tech/sources/tim-kellogg.md>)

Topics: [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Tool](<https://devfeed.tech/topics/tool.md>), [Parsing](<https://devfeed.tech/topics/parsing.md>), [Ollama](<https://devfeed.tech/topics/ollama.md>), [vllm](<https://devfeed.tech/topics/vllm.md>), [JSON](<https://devfeed.tech/topics/json.md>), [Reinforcement learning](<https://devfeed.tech/topics/reinforcement-learning.md>), [XML](<https://devfeed.tech/topics/xml.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [anthropic](<https://devfeed.tech/tags/anthropic.md>), [api](<https://devfeed.tech/tags/api.md>), [json](<https://devfeed.tech/tags/json.md>), [llm](<https://devfeed.tech/tags/llm.md>), [ollama](<https://devfeed.tech/tags/ollama.md>), [openai](<https://devfeed.tech/tags/openai.md>), [parsing](<https://devfeed.tech/tags/parsing.md>), [reinforcement-learning](<https://devfeed.tech/tags/reinforcement-learning.md>), [tools](<https://devfeed.tech/tags/tools.md>), [vllm](<https://devfeed.tech/tags/vllm.md>), [xml](<https://devfeed.tech/tags/xml.md>)

## AI overview

This article explains "inner loop agents," a design in which an LLM can parse and execute tool calls within the model-serving loop instead of returning control to a client. It discusses how software such as Ollama and vLLM handles model-specific tool-call formats and suggests that concurrent tool use during generation is a key motivation.

## Source excerpt

What if an LLM could use tools directly? In this post I discuss a potentially divergent view of agents, where agents are less like systems and more like LLMs specially trained to solve problems with a specific set of tools.