--- title: "open-dungeon vs agentflow" type: "comparison" canonical_url: "https://www.graphcanon.com/compare/newideas99-open-dungeon-vs-simonmesmith-agentflow" tools: ["newideas99-open-dungeon", "simonmesmith-agentflow"] --- # open-dungeon vs agentflow *GraphCanon updated Sep 20, 2026* ## Verdict Pick open-dungeon if open Dungeon stands out with its emphasis on full local processing of both text and image generation, utilizing Gemma 4 QAT via Ollama for interactive-fiction roleplaying experiences; pick agentflow if agentflow simplifies the creation of complex workflows for large language models through simple JSON configurations. [open-dungeon](https://github.com/newideas99/open-dungeon) reports 228 GitHub stars, 22 forks, and 1 open issues, last pushed Jul 6, 2026. [agentflow](https://github.com/simonmesmith/agentflow) has 320 stars, 27 forks, and 13 open issues, last pushed Aug 11, 2023. Figures are from public GitHub metadata via [open-dungeon's repository](https://github.com/newideas99/open-dungeon) and [agentflow's repository](https://github.com/simonmesmith/agentflow). | | [open-dungeon](/tools/newideas99-open-dungeon.md) | [agentflow](/tools/simonmesmith-agentflow.md) | | --- | --- | --- | | Tagline | easy-to-use fully local AI roleplay app with story and inline scene images generated on your machine | Complex LLM Workflows from Simple JSON | | Stars | 228 | 320 | | Forks | 22 | 27 | | Open issues | 1 | 13 | | Language | TypeScript | Python | | Adopt for | Open Dungeon stands out with its emphasis on full local processing of both text and image generation, utilizing Gemma 4 QAT via Ollama for interactive-fiction roleplaying experiences. | Agentflow simplifies the creation of complex workflows for large language models through simple JSON configurations. | | Persona | - | - | | Runtime | - | - | | License | MIT | MIT | | Categories | LLM Frameworks, Model Training | AI Agents, LLM Frameworks | ## Trust and health _Sourced signals - not a safety guarantee. No winner column._ | | [open-dungeon](/tools/newideas99-open-dungeon.md) | [agentflow](/tools/simonmesmith-agentflow.md) | | --- | --- | --- | | Maintenance | Steady (60%) | Dormant (18%) | | Days since push | 76d | 1100d | | Open issues (now) | 1 | 13 | | Stars delta | +16 (30d) | -1 (30d) | | Open issues delta | +1 (30d) | 0 (30d) | | Full report | [trust report](/tools/newideas99-open-dungeon/trust.md) | [trust report](/tools/simonmesmith-agentflow/trust.md) | ## Decision facts: open-dungeon - **Adopt for:** Open Dungeon stands out with its emphasis on full local processing of both text and image generation, utilizing Gemma 4 QAT via Ollama for interactive-fiction roleplaying experiences. ## Decision facts: agentflow - **Adopt for:** Agentflow simplifies the creation of complex workflows for large language models through simple JSON configurations. ## Choose when ### Choose open-dungeon if… - open-dungeon is primarily TypeScript; agentflow is Python. - Tags unique to open-dungeon: image-generation, interactive-fiction, local-llm, on-device-ai. - Also covers Model Training. - When you need to generate roleplay stories and inline scene images entirely offline without relying on any cloud services or API keys ### Choose agentflow if… - agentflow is primarily Python; open-dungeon is TypeScript. - Tags unique to agentflow: json, large-language-models, python, workflow-management. - Also covers AI Agents. - When you need to rapidly prototype LLM workflows with minimal coding via JSON configs ## When NOT to use open-dungeon - When you rely on continuous model updates delivered via the cloud since Open Dungeon processes everything locally and does not connect for updates - If your preferred environment is web-based with real-time collaboration features as Open Dungeon focuses purely on local execution, lacking online connectivity altogether ## When NOT to use agentflow - Avoid if requiring advanced customization that goes beyond basic JSON configurations - Not suitable for scenarios needing real-time dynamic changes in workflow setup during execution ## Common questions ### What is the difference between open-dungeon and agentflow? open-dungeon: easy-to-use fully local AI roleplay app with story and inline scene images generated on your machine. agentflow: Complex LLM Workflows from Simple JSON. See the comparison table for live GitHub stats and shared categories. ### When should I choose open-dungeon over agentflow? Choose open-dungeon over agentflow when open-dungeon is primarily TypeScript; agentflow is Python; Tags unique to open-dungeon: image-generation, interactive-fiction, local-llm, on-device-ai; Also covers Model Training; When you need to generate roleplay stories and inline scene images entirely offline without relying on any cloud services or API keys. ### When should I choose agentflow over open-dungeon? Choose agentflow over open-dungeon when agentflow is primarily Python; open-dungeon is TypeScript; Tags unique to agentflow: json, large-language-models, python, workflow-management; Also covers AI Agents; When you need to rapidly prototype LLM workflows with minimal coding via JSON configs. ### When should I avoid open-dungeon? When you rely on continuous model updates delivered via the cloud since Open Dungeon processes everything locally and does not connect for updates If your preferred environment is web-based with real-time collaboration features as Open Dungeon focuses purely on local execution, lacking online connectivity altogether ### When should I avoid agentflow? Avoid if requiring advanced customization that goes beyond basic JSON configurations Not suitable for scenarios needing real-time dynamic changes in workflow setup during execution ### Is open-dungeon or agentflow more popular on GitHub? agentflow has more GitHub stars (320 vs 228). Stars measure visibility, not whether either tool fits your constraints. ### Are open-dungeon and agentflow open source? Yes - both are open-source projects on GitHub (open-dungeon: MIT, agentflow: MIT). ### Where can I find alternatives to open-dungeon or agentflow? GraphCanon lists graph-backed alternatives at [open-dungeon alternatives](/tools/newideas99-open-dungeon/alternatives) and [agentflow alternatives](/tools/simonmesmith-agentflow/alternatives) ([open-dungeon markdown twin](/tools/newideas99-open-dungeon/alternatives.md), [agentflow markdown twin](/tools/simonmesmith-agentflow/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/newideas99-open-dungeon-vs-simonmesmith-agentflow.md) mirrors this page for agents and LLM crawlers, with the same stats table and FAQ answers. ### Which is better maintained, open-dungeon or agentflow? open-dungeon: Steady. agentflow: 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 open-dungeon and agentflow? GraphCanon publishes per-repo trust reports with dated maintenance, provenance, and scan summaries: [open-dungeon trust report](/tools/newideas99-open-dungeon/trust); [agentflow trust report](/tools/simonmesmith-agentflow/trust). --- **Machine-readable endpoints** - JSON: [`/api/graphcanon/graph?tool=newideas99-open-dungeon`](/api/graphcanon/graph?tool=newideas99-open-dungeon) - 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/_