Tendril
Loading lesson…
Hermes ships with a documented function-calling format. That makes it one of the few open-weight models you can wire into agent frameworks without months of prompting hacks.
For most production workloads, the model has to do something — call an API, query a database, write to a calendar. That is function calling. A model that handles it reliably is more valuable in agent stacks than one that writes prettier prose. Hermes was tuned with this in mind, and it shows.
Hermes uses a tool-use grammar that the model card documents in detail. The exact tags vary by version, but the pattern is consistent: the system prompt declares the available tools as JSON schemas, the model emits a structured tool-call block, your harness executes it and returns results, and the model continues. This is the same shape OpenAI and Anthropic use — Hermes just exposes it for an open-weight model.
Pattern (illustrative, not exact tags):
[SYSTEM]
You have access to the following tools:
search_inventory(query: string) -> list of items
send_email(to: string, subject: string, body: string)
[USER]
Find us 5 size 10 hiking boots and email John the list.
[MODEL]
<tool_call>
search_inventory({"query": "size 10 hiking boots"})
</tool_call>
[HARNESS RETURNS]
[ {sku: ..., name: ...}, ... ]
[MODEL]
<tool_call>
send_email({"to": "john@...", "subject": "Boots", "body": "..."})
</tool_call>The dance is the same in any tool-using model. Hermes documents the exact grammar in its model card.| Risk | What goes wrong | Mitigation |
|---|---|---|
| Schema drift | Model emits a slightly malformed tool call | Validate every call against the schema before executing |
| Hallucinated tool names | Model invents a tool you didn't declare | Allowlist tool names; refuse anything else |
| Argument hallucination | Model fills required fields with plausible-but-wrong values | Prompt to confirm before destructive actions |
| Multi-step planning | Model gets confused after 3-4 tool calls | Use a smaller plan-then-execute split or a larger model for planning |
The big idea: Hermes makes open-weight tool use practical. The work is in the harness — validate everything, allowlist always.
15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-hermes-function-calling-creators
What is the core idea behind "Hermes For Function Calling: Tool-Use Without OpenAI"?
Which term best describes a foundational idea in "Hermes For Function Calling: Tool-Use Without OpenAI"?
A learner studying Hermes For Function Calling: Tool-Use Without OpenAI would need to understand which concept?
Which of these is directly relevant to Hermes For Function Calling: Tool-Use Without OpenAI?
Which of the following is a key point about Hermes For Function Calling: Tool-Use Without OpenAI?
Which of these does NOT belong in a discussion of Hermes For Function Calling: Tool-Use Without OpenAI?
Which statement is accurate regarding Hermes For Function Calling: Tool-Use Without OpenAI?
Which of these does NOT belong in a discussion of Hermes For Function Calling: Tool-Use Without OpenAI?
What is the key insight about "Validate, don't trust" in the context of Hermes For Function Calling: Tool-Use Without OpenAI?
What is the key insight about "Don't expose destructive tools to small models" in the context of Hermes For Function Calling: Tool-Use Without OpenAI?
What is the key insight about "From the community" in the context of Hermes For Function Calling: Tool-Use Without OpenAI?
Which statement accurately describes an aspect of Hermes For Function Calling: Tool-Use Without OpenAI?
What does working with Hermes For Function Calling: Tool-Use Without OpenAI typically involve?
Which of the following is true about Hermes For Function Calling: Tool-Use Without OpenAI?
Which best describes the scope of "Hermes For Function Calling: Tool-Use Without OpenAI"?