Comparative Analysis of Agentic Orchestration Layers: Personalization, Planning, Search, MCP Integration, and Reasoning Across Twelve Open-Source Frameworks

Authors

  • Ankur Aggarwal

Keywords:

Agentic AI, LLM Orchestration, Retrieval-Augmented Generation, Model Context Protocol, Multi-Agent Systems, Open-Source Frameworks, Reasoning Patterns

Abstract

INTRODUCTION: The rapid growth of LLM-powered agentic systems has produced a fragmented ecosystem of open-source orchestration frameworks addressing agent memory, planning, retrieval, and reasoning.

OBJECTIVES: To comparatively evaluate twelve leading open-source orchestration frameworks across dimensions critical to production deployment.

METHODS: Twelve frameworks were evaluated across five dimensions: personalization, memory architecture, planning paradigms, retrieval integration, and Model Context Protocol (MCP) adoption depth.

RESULTS: Findings include a personalization spectrum from stateless to memory-first designs, a planning-control trade-off as the primary selection axis, and near-universal MCP adoption where integration depth is the differentiating factor.

CONCLUSION: Framework selection affects reliability, scalability, and production viability; the five-dimensional evaluation provides actionable heuristics for matching framework architecture to deployment context.

Downloads

Download data is not yet available.

References

Wang L, Ma C, Feng X, Zhang Z, Yang H, Zhang J, et al. A survey on large language model based autonomous agents. Front Comput Sci. 2024;18:186345. doi:10.1007/s11704-024-40231-1

Hong S, Zhuge M, Chen J, Zheng X, Cheng Y, Zhang C, et al. MetaGPT: meta programming for a multi-agent collaborative framework. In: Proceedings of the 12th International Conference on Learning Representations (ICLR 2024); 2024 May 7–11; Vienna, Austria. OpenReview.net; 2024. Available from: https://openreview.net/forum?id=VtmBAGCN7o

Khattab O, Singhvi A, Maheshwari P, Zhang Z, Santhanam K, Vardhamanan S, et al. DSPy: compiling declarative language model calls into self-improving pipelines. In: Proceedings of the 12th International Conference on Learning Representations (ICLR 2024); 2024 May 7–11; Vienna, Austria. OpenReview.net; 2024. Available from: https://openreview.net/forum?id=sY5N0zY5Od

Anthropic. Introducing the Model Context Protocol [Internet]. Anthropic; 2024 Nov [cited 2025 Jun 3]. Available from: https://www.anthropic.com/news/model-context-protocol

Yao S, Zhao J, Yu D, Du N, Shafran I, Narasimhan K, et al. ReAct: synergizing reasoning and acting in language models. In: Proceedings of the 11th International Conference on Learning Representations (ICLR 2023); 2023 May 1–5; Kigali, Rwanda. OpenReview.net; 2023. Available from: https://openreview.net/forum?id=WE_vluYUL-X

Shinn N, Cassano F, Berman E, Gopinath A, Narasimhan K, Yao S. A. Reflexion: an autonomous agent with dynamic memory and self-reflection. arXiv:230311366 [cs] [Internet]. 2023 Mar 20; Available from: https://arxiv.org/abs/2303.11366

LangChain. LangGraph [Internet]. www.langchain.com. Available from: https://www.langchain.com/langgraph

Mintlify. CrewAI [Internet]. Crewai.com. CrewAI; 2024 [cited 2026 Jun 3]. Available from: https://docs.crewai.com/concepts/memory

OpenAI Agents SDK [Internet]. Github.io. 2025. Available from: https://openai.github.io/openai-agents-python

Agent Development Kit [Internet]. Google Cloud Documentation. 2026. Available from: https://docs.cloud.google.com/gemini-enterprise-agent-platform/build/adk

LastMile AI. mcp-agent: build powerful agents using Model Context Protocol [Internet]. GitHub; 2025 [cited 2025 Jun 3]. Available from: https://github.com/lastmile-ai/mcp-agent

deepset. Haystack: the open-source framework for building production-ready LLM applications [Internet]. GitHub; 2025 [cited 2026 Jun 3]. Available from: https://github.com/deepset-ai/haystack

LlamaIndex. Agentic RAG with LlamaIndex [Internet]. LlamaIndex Blog; 2024. Available from: https://www.llamaindex.ai/blog/agentic-rag-with-llamaindex-2721b8a49ff6

Microsoft. Microsoft Agent Framework overview [Internet]. Microsoft Learn; 2026 [cited 2026 Jun 3]. Available from: https://learn.microsoft.com/en-us/agent-framework/overview

Meta AI. Model cards and prompt formats: Llama 4 [Internet]. Meta AI; [cited 2026 Jun 3]. Available from: https://www.llama.com/docs/model-cards-and-prompt-formats/llama4/

Xu J, Li Z, Chen W, Wang Q, Gao X, Cai Q, et al. On-device language models: a comprehensive review. arXiv [Internet]. 2024 Aug. doi:10.48550/arXiv.2409.00088

Artificial Analysis. Independent analysis of AI models and API providers [Internet]. Artificial Analysis; 2026 [cited 2026 Jun 3]. Available from: https://artificialanalysis.ai

Grattafiori A, Dubey A, Jauhri A, Pandey A, Kadian A, Al-Dahle A, et al. The Llama 3 herd of models. arXiv [Internet]. 2024 Jul. Report No.: arXiv:2407.21783. doi:10.48550/arXiv.2407.21783

Model Context Protocol. MCP specification transports [Internet]. Model Context Protocol; 2025 Mar [cited 2025 Jun 3]. Available from: https://modelcontextprotocol.io/specification/2025-11-25/basic/transports

Patil SG, Mao H, Yan F, Ji CC, Suresh V, Stoica I, et al. The Berkeley Function Calling Leaderboard (BFCL): from tool use to agentic evaluation of large language models. In: Proceedings of the 42nd International Conference on Machine Learning (ICML 2025); 2025; Vancouver, Canada. PMLR; 2025. vol. 267. p. 48371–48392. Available from: https://proceedings.mlr.press/v267/patil25a.html

Abou Ali M, Dornaika F. Agentic AI: a comprehensive survey of architectures, applications, and future directions. arXiv [Internet]. 2025 Oct [cited 2025 Jun 3]. Report No.: arXiv:2510.25445. doi:10.48550/arXiv.2510.25445

Downloads

Published

10.07.2026

How to Cite

Ankur Aggarwal. (2026). Comparative Analysis of Agentic Orchestration Layers: Personalization, Planning, Search, MCP Integration, and Reasoning Across Twelve Open-Source Frameworks. International Journal of Intelligent Systems and Applications in Engineering, 14(1s), 1965–1976. Retrieved from https://www.ijisae.org/index.php/IJISAE/article/view/8451

Issue

Vol. 14 No. 1s (2026)

Section

Research Article

License

Creative Commons License

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

All papers should be submitted electronically. All submitted manuscripts must be original work that is not under submission at another journal or under consideration for publication in another form, such as a monograph or chapter of a book. Authors of submitted papers are obligated not to submit their paper for publication elsewhere until an editorial decision is rendered on their submission. Further, authors of accepted papers are prohibited from publishing the results in other publications that appear before the paper is published in the Journal unless they receive approval for doing so from the Editor-In-Chief.

IJISAE open access articles are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. This license lets the audience to give appropriate credit, provide a link to the license, and indicate if changes were made and if they remix, transform, or build upon the material, they must distribute contributions under the same license as the original.