⚡
AgentForge

Features

Everything you need to deploy autonomous agents securely at scale.

Security

Capability Tokens

HMAC-signed tokens specify exactly what each agent can access: files, domains, token budget, timeout.

Filesystem ACLs

Glob patterns restrict file access. Agents cannot escape their allowed directories.

Domain Whitelist

Network calls only to approved domains. No unrestricted internet access.

Token Budget

Set max tokens per session. Prevent runaway costs and infinite loops.

Timeout Enforcement

All agents have a hard timeout. No hung processes eating resources.

Audit Trail

Complete logging of every tool call, ACL check, and security event.

Capabilities

File I/O

Read, write, delete, list files—all ACL-protected. Handle JSON, markdown, text, code.

Web Access

HTTP requests and Google Search integration. Domain-whitelist protected.

Memory Store (MeMex RAG)

Persistent memory with Git versioning, full-text search, and semantic retrieval.

LLM Integration

Call any LLM provider. Direct chat or embedding generation.

Content Generation

DALL-E, Stability AI, Runway, Synthesia, Mermaid diagrams, AI music.

Development Tools

Shell execution (safe), code review, API design, statistical analysis.

Operations

Inter-Agent Communication

CSP message bus with topic subscriptions. Agents coordinate without central broker.

Pipeline Orchestration

DAG-based pipelines. Stages run in parallel or sequentially. Full dependency management.

Real-time Dashboard

Web UI to spawn agents, monitor status, view logs, manage memory, visualize pipelines.

Cost Tracking

Real-time token usage and cost per agent. Set budgets and alerts.

Memory Management

Automatic session compaction, Git versioning, full-text indexing.

Monitoring & Metrics

Prometheus metrics, structured logging, Grafana integration ready.

Ready to Get Started?

Read the docs, try the open-source version, or contact us for enterprise support.

View Documentation