murmur
★ 20
vs
OctoForge
★ 22
murmur vs OctoForge
murmur: Murmur is a secure, end-to-end encrypted multi-agent communication bus that enables AI models like Claude and GPT to talk directly to each other. It eliminates the need for human relay, allowing for autonomous collaboration on complex tasks through features like `murmur_request` for send-and-wait messaging.; OctoForge: OctoForge is a self-hosted, multi-user agent platform designed for teams, enabling a single assistant to learn skills, remember context, and integrate with internal APIs. It supports concurrent interactions with multiple users and tasks, while a single process can host various organizations, each with isolated resources.
01
TL;DR
murmurMulti-Agent · ★ 20
Autonomous Code Reviews: AI agents can collaborate on reviewing pull requests and providing feedback without human intervention.
OctoForgeMulti-Agent · ★ 22
Internal Knowledge Assistant: Providing a company-wide AI assistant that learns from internal documentation and answers employee questions.
02
Side-by-Side Comparison
Field
murmurOctoForgeCategory
Multi-Agent
Multi-Agent
Stars
★ 20
★ 22
License
MIT
NOASSERTION
Install Command
git clone https://github.com/alexfrmn/murmur.git && cd murmur && git checkout --detach v2.10.0 && npm ci && npm run build
git clone https://github.com/dmirain/OctoForge && cd OctoForge && make quickstart
Updated
2d ago
1mo ago
Open Source
Yes
Yes
Tags
Multi-Agent Communication, E2E Encryption, NATS Messaging
Multi-Agent Platform, Self-Hosted, Python
03
Features
murmur01E2E Encryption & Digital Signatures: Uses X25519 + XChaCha20-Poly1305 AEAD for encryption and Ed25519 for message authentication, ensuring secure communication.
02`murmur_request` for Autonomous Workflows: A send-and-wait mechanism that automatically polls for replies, enabling fully autonomous, overnight agent collaboration without human intervention.
03At-Least-Once Delivery with SQLite Outbox: Guarantees message delivery through a persistent SQLite outbox with ACK correlation and includes a dead-letter queue for failed messages.
04MCP Server for AI Client Integration: Exposes a JSON-RPC MCP server with 7 tools, facilitating integration with any MCP-compatible AI client, including native wake for Claude and Codex.
05Scoped Channels & Session Affinity: Ensures that for an addressed conversation, only the owning session of the addressed agent responds, preventing competing threads and guaranteeing exactly one message emit.
OctoForge01New capabilities arrive without a deploy: Skills, knowledge, and API integrations are stored in the database, allowing the agent to learn and adapt without code changes or restarts.
02A dialog is an actor, not a request handler: Conversations are durable, concurrent "exchanges" that can pause for user input or run background tasks, preventing waits.
03Secrets are structurally out of reach of the model: API keys and tokens are stored encrypted and accessed via one-time links, ensuring credentials never pass through the LLM context.
04Multi-user is the schema, not a wrapper: User and tenant isolation (ownership, access) is built into the database schema from the ground up, not added as an afterthought.
05Big API responses live in the database, not in the context: Large API payloads are stored and processed in Postgres as "collections," providing the model with summarized insights rather than raw data.
04
Use Cases
murmur↳Autonomous Code Reviews: AI agents can collaborate on reviewing pull requests and providing feedback without human intervention.
↳Collaborative Research Tasks: Multiple AI agents can work together on research, sharing context and findings through encrypted channels.
↳Automated Architectural Decisions: Agents can autonomously discuss and make architectural decisions, leveraging different models' strengths.
↳Inter-Organizational Agent Communication: Facilitates secure, encrypted communication between AI agents across different machines and organizations.
OctoForge↳Internal Knowledge Assistant: Providing a company-wide AI assistant that learns from internal documentation and answers employee questions.
↳Automated Workflow Execution: Scheduling tasks, calling internal APIs with secured credentials, and performing background work based on user requests.
↳Multi-Tenant Bot Hosting: Hosting multiple isolated AI assistants for different departments or client organizations within a single self-hosted instance.
↳Secure API Gateway for LLMs: Allowing LLMs to interact with sensitive internal APIs without directly exposing credentials to the model.
↳Enhanced Customer Support Bot: Deploying a bot that can access knowledge bases, query user-specific data, and manage support tickets.
FAQ
FAQ
What is the difference between murmur and OctoForge?
Both murmur and OctoForge are in the Multi-Agent category. murmur has 20 stars, while OctoForge has 22 stars.
Which is better, murmur or OctoForge?
The best choice depends on your use case. Choose murmur if Autonomous Code Reviews: AI agents can collaborate on reviewing pull requests and providing feedback without human intervention., and OctoForge if Internal Knowledge Assistant: Providing a company-wide AI assistant that learns from internal documentation and answers employee questions..
Is murmur free or open source?
Yes, murmur is open source on GitHub (MIT).
Is OctoForge free or open source?
Yes, OctoForge is open source on GitHub (NOASSERTION).
Cloud Infrastructure
AdDeploy on DigitalOcean
Host your MCP servers, Docker containers, and AI agents with $200 free credit over 60 days.
→