--- title: "mcp-agent vs agency" type: "comparison" canonical_url: "https://www.graphcanon.com/compare/lastmile-ai-mcp-agent-vs-neurocult-agency" tools: ["lastmile-ai-mcp-agent", "neurocult-agency"] --- # mcp-agent vs agency *GraphCanon updated Aug 26, 2026* ## Verdict Pick mcp-agent if allows creation of efficient AI agents via Model Context Protocol with workflow patterns; pick agency if agency is a Go-idiomatic library aimed at developers looking to work with Large Language Models and generative AI techniques. [mcp-agent](https://github.com/lastmile-ai/mcp-agent) reports 8.5k GitHub stars, 879 forks, and 134 open issues, last pushed Jan 25, 2026. [agency](https://github.com/neurocult/agency) has 514 stars, 36 forks, and 4 open issues, last pushed Jan 8, 2025. Figures are from public GitHub metadata via [mcp-agent's repository](https://github.com/lastmile-ai/mcp-agent) and [agency's repository](https://github.com/neurocult/agency). | | [mcp-agent](/tools/lastmile-ai-mcp-agent.md) | [agency](/tools/neurocult-agency.md) | | --- | --- | --- | | Tagline | Build effective AI agents with Model Context Protocol. | Library for exploring Large Language Models and generative AI in Go | | Stars | 8,518 | 514 | | Forks | 879 | 36 | | Open issues | 134 | 4 | | Language | Python | Go | | Adopt for | Allows creation of efficient AI agents via Model Context Protocol with workflow patterns | Agency is a Go-idiomatic library aimed at developers looking to work with Large Language Models and generative AI techniques. | | Persona | - | - | | Runtime | - | - | | License | Apache-2.0 | MIT | | Categories | AI Agents | AI Agents, LLM Frameworks | ## Trust and health _Sourced signals - not a safety guarantee. No winner column._ | | [mcp-agent](/tools/lastmile-ai-mcp-agent.md) | [agency](/tools/neurocult-agency.md) | | --- | --- | --- | | Maintenance | Slowing (36%) | Dormant (18%) | | Days since push | 212d | 589d | | Open issues (now) | 134 | 4 | | Stars delta | +40 (30d) | +2 (30d) | | Open issues delta | -6 (30d) | 0 (30d) | | Full report | [trust report](/tools/lastmile-ai-mcp-agent/trust.md) | [trust report](/tools/neurocult-agency/trust.md) | ## Decision facts: mcp-agent - **Adopt for:** Allows creation of efficient AI agents via Model Context Protocol with workflow patterns ## Decision facts: agency - **Pricing:** freemium - Agency is available under MIT license and is free for use, modification, and distribution. Additional services or integrations might involve costs not outlined within the repository. - **Adopt for:** Agency is a Go-idiomatic library aimed at developers looking to work with Large Language Models and generative AI techniques. ## Choose when ### Choose mcp-agent if… - mcp-agent is primarily Python; agency is Go. - License: mcp-agent is Apache-2.0, agency is MIT. - Tags unique to mcp-agent: ai-agents, llm, mcp, model-context-protocol. - Need to implement the Model Context Protocol specifically in your agent development ### Choose agency if… - agency is primarily Go; mcp-agent is Python. - License: agency is MIT, mcp-agent is Apache-2.0. - Pricing: Agency is available under MIT license and is free for use, modification, and distribution. Additional services or integrations might involve costs not outlined within the repository.. - Tags unique to agency: generative-ai, go, language-models, neural-networks. - Also covers LLM Frameworks. - If you're proficient in Go and want to implement LLMs within a familiar ecosystem, consider agency. ## When NOT to use mcp-agent - Require a more customizable approach beyond pre-defined protocol and workflows - Project demands specific AI frameworks unsupported by Model Context Protocol directly ## When NOT to use agency - Avoid agency if your primary programming expertise lies outside of the Go ecosystem. - Not recommended if you require real-time performance characteristics that surpass what typical LLM exploration libraries can provide. ## Common questions ### What is the difference between mcp-agent and agency? mcp-agent: Build effective AI agents with Model Context Protocol.. agency: Library for exploring Large Language Models and generative AI in Go. See the comparison table for live GitHub stats and shared categories. ### When should I choose mcp-agent over agency? Choose mcp-agent over agency when mcp-agent is primarily Python; agency is Go; License: mcp-agent is Apache-2.0, agency is MIT; Tags unique to mcp-agent: ai-agents, llm, mcp, model-context-protocol; Need to implement the Model Context Protocol specifically in your agent development. ### When should I choose agency over mcp-agent? Choose agency over mcp-agent when agency is primarily Go; mcp-agent is Python; License: agency is MIT, mcp-agent is Apache-2.0; Pricing: Agency is available under MIT license and is free for use, modification, and distribution. Additional services or integrations might involve costs not outlined within the repository.; Tags unique to agency: generative-ai, go, language-models, neural-networks; Also covers LLM Frameworks; If you're proficient in Go and want to implement LLMs within a familiar ecosystem, consider agency. ### When should I avoid mcp-agent? Require a more customizable approach beyond pre-defined protocol and workflows Project demands specific AI frameworks unsupported by Model Context Protocol directly ### When should I avoid agency? Avoid agency if your primary programming expertise lies outside of the Go ecosystem. Not recommended if you require real-time performance characteristics that surpass what typical LLM exploration libraries can provide. ### Is mcp-agent or agency more popular on GitHub? mcp-agent has more GitHub stars (8,518 vs 514). Stars measure visibility, not whether either tool fits your constraints. ### Are mcp-agent and agency open source? Yes - both are open-source projects on GitHub (mcp-agent: Apache-2.0, agency: MIT). ### Where can I find alternatives to mcp-agent or agency? GraphCanon lists graph-backed alternatives at [mcp-agent alternatives](/tools/lastmile-ai-mcp-agent/alternatives) and [agency alternatives](/tools/neurocult-agency/alternatives) ([mcp-agent markdown twin](/tools/lastmile-ai-mcp-agent/alternatives.md), [agency markdown twin](/tools/neurocult-agency/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/lastmile-ai-mcp-agent-vs-neurocult-agency.md) mirrors this page for agents and LLM crawlers, with the same stats table and FAQ answers. ### Which is better maintained, mcp-agent or agency? mcp-agent: Slowing. agency: Dormant. 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 mcp-agent and agency? GraphCanon publishes per-repo trust reports with dated maintenance, provenance, and scan summaries: [mcp-agent trust report](/tools/lastmile-ai-mcp-agent/trust); [agency trust report](/tools/neurocult-agency/trust). --- **Machine-readable endpoints** - JSON: [`/api/graphcanon/graph?tool=lastmile-ai-mcp-agent`](/api/graphcanon/graph?tool=lastmile-ai-mcp-agent) - 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/_