v0.1.0 — Released March 2026

Define AI_
workflows as YAML.
Deploy like infrastructure.

Single binary. Postgres only. Bring your own keys. 9 connectors, infinite plugins.

9 built-in connectors 17 example workflows Source-available (BSL 1.1)

fetch-and-summarize.yaml

name: fetch-and-summarize
description: Fetch data from an API and summarize it

inputs:
  url:
    type: string
    description: URL to fetch

steps:
  - name: fetch-data
    action: http/request
    params:
      method: GET
      url: "{{ inputs.url }}"

  - name: summarize
    action: ai/completion
    credential: my-openai-key
    params:
      model: gpt-4o
      prompt: "Summarize: {{ steps['fetch-data'].output.body }}"
      output_schema:
        type: object
        properties:
          summary:
            type: string
          key_points:
            type: array

Why Mantle?

IaC Lifecycle

validate, plan, apply, run. The same workflow you use for Terraform, applied to workflow automation. Pin executions to immutable versions. Diff before you deploy.

Single Binary

One Go binary. One Postgres database. No message queues, no worker fleets, no cluster topology. Deploy anywhere containers run.

BYOK

Your API keys live in your database, encrypted with your encryption key. Mantle never proxies through a hosted service. OpenAI, Anthropic, Bedrock, Azure, or self-hosted.

AI Tool Use

Multi-turn function calling out of the box. The LLM requests tools, Mantle executes them via connectors, feeds results back. Crash recovery included.

How It Works

Write

Define your workflow as a YAML file

Validate

Check structure offline, run in CI

Apply

Store an immutable version in Postgres

Run

Execute with checkpoint-and-resume

Terminal

$ mantle validate examples/fetch-and-summarize.yaml
fetch-and-summarize.yaml: valid
 
$ mantle plan examples/fetch-and-summarize.yaml
+ fetch-and-summarize (new workflow, version 1)
 
$ mantle apply examples/fetch-and-summarize.yaml
Applied fetch-and-summarize version 1
 
$ mantle run fetch-and-summarize --input url=https://api.example.com/data
Running fetch-and-summarize (version 1)...
Execution a1b2c3d4: completed
  fetch-data: completed (0.8s)
  summarize:  completed (2.1s)

Built-in Connectors

7 connectors with 9 actions ship with the binary — plus a plugin system for anything else.

HTTP

REST APIs, webhooks, any HTTP endpoint

AI (OpenAI)

Chat completions, structured output, tool use

AI (Bedrock)

AWS Bedrock models with region routing

Slack

Send messages, read channel history

Email

Send via SMTP, plaintext and HTML

Postgres

Parameterized SQL against external databases

S3

Put, get, list objects (S3-compatible)

+ Plugins

Any executable. JSON in, JSON out. Python, Go, Bash.

How Mantle Compares

	Mantle	Temporal	n8n / Zapier	LangChain	Prefect / Airflow
Primary use case	AI workflow automation	Distributed transactions	SaaS integration	LLM app development	Data pipelines
Workflow format	YAML + CEL	Go / Java SDK	Visual canvas	Python code	Python code
Deployment	Single binary + Postgres	Multi-service cluster	Self-hosted or cloud	Library in your app	Python platform
AI/LLM support	First-class (built-in)	Build it yourself	Partial (AI nodes)	First-class (library)	Build it yourself
Checkpointing	Built-in	Built-in	Partial	None	Built-in
Secrets management	Built-in, encrypted	External	Built-in	External	External
Version control	Git-native (IaC)	Code in repos	Database-stored	Code in repos	Code in repos
Target user	Platform engineers	Backend engineers	Non-technical users	Python developers	Data engineers
Operational complexity	Low	High	Medium	N/A (library)	Medium-High

Mantle is early (v0.1.0). Temporal, Airflow, and LangChain have years of production hardening and larger ecosystems. Choose the tool that fits your use case and team.

Built for Production

Enterprise features ship in the same binary. No upgrade tiers, no feature gates.

Multi-tenancy & RBAC

Teams, users, roles (admin / team_owner / operator). API key authentication with hashed storage.

OIDC / SSO

Token-sniffing middleware supports API keys and OIDC tokens. Integrate with your existing identity provider.

Encrypted Credentials

AES-256-GCM encryption at rest. Cloud secret backends: AWS Secrets Manager, GCP Secret Manager, Azure Key Vault.

Audit Trail

Every state-changing operation emits an immutable audit event to Postgres. Query with mantle audit.

Prometheus Metrics

Workflow execution counts, step durations, connector latencies, active executions gauge. Scrape /metrics in server mode.

Helm Chart

Production-ready Kubernetes deployment. PDB, migration job, startup probes, security contexts, ServiceMonitor.

Rate Limiting

Protect external APIs and control execution throughput.

Data Retention

Configure retention policies for execution history and audit events.

Multi-arch Images

amd64 + arm64 Docker images published to GHCR. govulncheck, gosec, and Trivy scanning in CI.

Get Running in 5 Minutes

From zero to running workflow in four commands.

1 Install

$ go install github.com/dvflw/mantle/cmd/mantle@latest

2 Start Postgres and initialize

$ docker compose up -d
$ mantle init

3 Apply your first workflow

$ mantle apply examples/hello-world.yaml
# Applied hello-world version 1

4 Run it

$ mantle run hello-world
# Running hello-world (version 1)...
# Execution a1b2c3d4: completed
#   fetch: completed (1.0s)

17 example workflows included. HTTP, AI, Slack, Postgres, S3, parallel execution, cron triggers, webhooks, and multi-turn tool use.

Read the Getting Started guide

Define AI_ workflows as YAML. Deploy like infrastructure.