---
title: "gateway vs guidance"
type: "comparison"
canonical_url: "https://www.graphcanon.com/compare/adaline-gateway-vs-guidance-ai-guidance"
tools: ["adaline-gateway", "guidance-ai-guidance"]
---

# gateway vs guidance

*GraphCanon updated Sep 30, 2026*

## Verdict

Pick gateway if adaline Gateway is a production-grade, fully local Super SDK that supports over 300 LLMs with a unified interface, including features like batching, retries, caching, and OpenTelemetry integration; pick guidance if guidance is a specialized tool written in Jupyter Notebooks that provides a unique language to control large language models (LLMs) across multiple backends such as Transformers, llama.cpp.

[gateway](https://www.adaline.ai/docs) reports 607 GitHub stars, 26 forks, and 2 open issues, last pushed Jul 29, 2026. [guidance](https://github.com/guidance-ai/guidance) has 22k stars, 1.2k forks, and 323 open issues, last pushed May 21, 2026. Figures are from public GitHub metadata via [gateway's repository](https://github.com/adaline/gateway) and [guidance's repository](https://github.com/guidance-ai/guidance).

| | [gateway](/tools/adaline-gateway.md) | [guidance](/tools/guidance-ai-guidance.md) |
| --- | --- | --- |
| Tagline | Super SDK for calling over 300 LLMs with a unified interface | A guidance language for controlling large language models. |
| Stars | 607 | 21,739 |
| Forks | 26 | 1,203 |
| Open issues | 2 | 323 |
| Language | TypeScript | Jupyter Notebook |
| Adopt for | Adaline Gateway is a production-grade, fully local Super SDK that supports over 300 LLMs with a unified interface, including features like batching, retries, caching, and OpenTelemetry integration. | Guidance is a specialized tool written in Jupyter Notebooks that provides a unique language to control large language models (LLMs) across multiple backends such as Transformers, llama.cpp, and OpenAI. It's open-source,轻 |
| Persona | - | - |
| Runtime | - | - |
| License | MIT | MIT |
| Categories | Inference & Serving, LLM Frameworks | Inference & Serving, LLM Frameworks |

## Trust and health

_Sourced signals - not a safety guarantee. No winner column._

| | [gateway](/tools/adaline-gateway.md) | [guidance](/tools/guidance-ai-guidance.md) |
| --- | --- | --- |
| Maintenance | Steady (60%) | Slowing (36%) |
| Days since push | 52d | 108d |
| Open issues (now) | 2 | 323 |
| Stars delta | +2 (30d) | +33 (30d) |
| Open issues delta | 0 (30d) | +7 (30d) |
| Full report | [trust report](/tools/adaline-gateway/trust.md) | [trust report](/tools/guidance-ai-guidance/trust.md) |

## Decision facts: gateway

- **Adopt for:** Adaline Gateway is a production-grade, fully local Super SDK that supports over 300 LLMs with a unified interface, including features like batching, retries, caching, and OpenTelemetry integration.

## Decision facts: guidance

- **Adopt for:** Guidance is a specialized tool written in Jupyter Notebooks that provides a unique language to control large language models (LLMs) across multiple backends such as Transformers, llama.cpp, and OpenAI. It's open-source,轻

## Choose when

### Choose gateway if…

- gateway is primarily TypeScript; guidance is Jupyter Notebook.
- Tags unique to gateway: ai, ai-agents, anthropic, language-model.
- When you need a fully local SDK that supports more than 300 LLMs with a unified interface.

### Choose guidance if…

- guidance is primarily Jupyter Notebook; gateway is TypeScript.
- Tags unique to guidance: backend support, control language, language models, pip-installable.
- When you need a specific language to finely control various LLM backends including Transformers, llama.cpp, and OpenAI

## When NOT to use gateway

- If you are looking for a cloud-based service, as Adaline Gateway is fully local and not a proxy.
- When you need support for a specific LLM that is not among the 300+ supported by Adaline Gateway.
- If you do not require or want the complexity of a fully featured SDK with extensive customization options.
- When you prefer a simpler, less extensible solution for calling LLMs.

## When NOT to use guidance

- When your project is strictly confined to using only one type of backend which you can manage without a specialized control language
- If your development environment does not support or prefer Jupyter Notebooks, Guidance may not be the best choice

## Common questions

### What is the difference between gateway and guidance?

gateway: Super SDK for calling over 300 LLMs with a unified interface. guidance: A guidance language for controlling large language models.. See the comparison table for live GitHub stats and shared categories.

### When should I choose gateway over guidance?

Choose gateway over guidance when gateway is primarily TypeScript; guidance is Jupyter Notebook; Tags unique to gateway: ai, ai-agents, anthropic, language-model; When you need a fully local SDK that supports more than 300 LLMs with a unified interface.

### When should I choose guidance over gateway?

Choose guidance over gateway when guidance is primarily Jupyter Notebook; gateway is TypeScript; Tags unique to guidance: backend support, control language, language models, pip-installable; When you need a specific language to finely control various LLM backends including Transformers, llama.cpp, and OpenAI.

### When should I avoid gateway?

If you are looking for a cloud-based service, as Adaline Gateway is fully local and not a proxy. When you need support for a specific LLM that is not among the 300+ supported by Adaline Gateway. If you do not require or want the complexity of a fully featured SDK with extensive customization options. When you prefer a simpler, less extensible solution for calling LLMs.

### When should I avoid guidance?

When your project is strictly confined to using only one type of backend which you can manage without a specialized control language If your development environment does not support or prefer Jupyter Notebooks, Guidance may not be the best choice

### Is gateway or guidance more popular on GitHub?

guidance has more GitHub stars (21,739 vs 607). Stars measure visibility, not whether either tool fits your constraints.

### Are gateway and guidance open source?

Yes - both are open-source projects on GitHub (gateway: MIT, guidance: MIT).

### Where can I find alternatives to gateway or guidance?

GraphCanon lists graph-backed alternatives at [gateway alternatives](/tools/adaline-gateway/alternatives) and [guidance alternatives](/tools/guidance-ai-guidance/alternatives) ([gateway markdown twin](/tools/adaline-gateway/alternatives.md), [guidance markdown twin](/tools/guidance-ai-guidance/alternatives.md)), ranked by typed relationship edges rather than popularity votes.

### Is there a machine-readable version of this comparison?

Yes. The markdown twin at [this comparison](/compare/adaline-gateway-vs-guidance-ai-guidance.md) mirrors this page for agents and LLM crawlers, with the same stats table and FAQ answers.

### Which is better maintained, gateway or guidance?

gateway: Steady. guidance: Slowing. Compare maintenance labels, days since push, and release cadence in the trust section below - stars alone do not measure maintenance.

### Where are the full trust reports for gateway and guidance?

GraphCanon publishes per-repo trust reports with dated maintenance, provenance, and scan summaries: [gateway trust report](/tools/adaline-gateway/trust); [guidance trust report](/tools/guidance-ai-guidance/trust).

---

**Machine-readable endpoints**

- JSON: [`/api/graphcanon/graph?tool=adaline-gateway`](/api/graphcanon/graph?tool=adaline-gateway)
- LLM index: [/llms.txt](/llms.txt)
- Full corpus: [/llms-full.txt](/llms-full.txt)

_GraphCanon - The knowledge graph for AI development. https://www.graphcanon.com/_
