Skip to main content

Compare FST

FST does not replace your agent stack. It makes it trustworthy.

Agent frameworks run agents. Workflow tools move work between systems. MCP exposes tools, resources, and prompts. Chat gateways connect users to assistants. FST controls the process the agent is working inside.

Try the quickstartSee pricing
Agents comparing work across systems with a trusted approval path
Capability is not authorityAn agent may be able to call a tool, draft a change, or request a deployment. FST decides whether that step is allowed inside the active process.
Keep the tools that do the workAgent frameworks, workflow tools, MCP servers, and chat gateways remain useful. FST gives them controlled runs, gates, approvals, and evidence.
Trust comes before the effectObservability can explain what happened. FST checks scope, artifacts, approvals, and preflight before protected work is allowed to count.

FST is the control point, not the worker.

Use whatever agent or workflow system you like. FST gives it a persistent, enforceable process to work inside.

Agent frameworks ask: how should the task be done?FST asks: is this process step valid?
Workflow tools ask: what runs next?FST asks: what must be true before progress counts?
MCP exposes: which tools are available?FST checks: is this tool use authorized in this run?
Observability asks: what happened?FST asks: was it allowed before it happened?

How FST complements the tools you already use

FST does not compete on model quality, workflow integrations, chat channels, or generic tool exposure. It makes those systems more useful for process-sensitive work by separating capability from authority.

Plain agents

Codex, Claude, Gemini, Kimi, opencode

Plain agents reason, draft, inspect systems, write code, and request tool calls.

How FST makes it stronger: FST makes their work more trustworthy by giving them official run state, scoped authority, required gates, typed artifacts, approval checks, and replayable evidence.

Read the detailed post

LangGraph

Stateful agent graphs

LangGraph gives agents durable graph execution, checkpoints, and structured paths through complex tasks.

How FST makes it stronger: FST makes LangGraph stronger by acting as the process authority at graph checkpoints. A graph node can ask FST what is missing before the graph advances or touches a protected system.

Read the detailed post

CrewAI

Multi-agent crews and flows

CrewAI coordinates agents, tasks, roles, and flows so multiple agents can collaborate on work.

How FST makes it stronger: FST makes crews safer by providing shared process state and gatekeeping across agents. A crew can submit work, while FST decides whether the evidence satisfies the process.

Read the detailed post

n8n and workflow tools

Triggers, integrations, handoffs, app actions

Workflow tools move data between systems, react to triggers, run automations, and execute app actions.

How FST makes it stronger: FST makes workflow automation more trustworthy by checking whether an agent-generated artifact, approval, or protected effect is ready before the workflow executes the action.

Read the detailed post

MCP and tool layers

Tools, resources, prompts, adapters

MCP lets agents discover and call tools through a common interface.

How FST makes it stronger: FST can expose MCP tools, but the value is behind the interface: profile lookup, gate evaluation, scope authorization, approval validation, evidence writes, and replay.

Read the detailed post

Chat gateways

OpenClaw, chat channels, assistant gateways

Chat gateways connect users to assistants through Slack, Teams, Telegram, WhatsApp, or other channels.

How FST makes it stronger: FST makes chat-driven work trustworthy by preventing a message from automatically becoming authority. Messages can become requests or approval candidates only when the process validates them.

Read the detailed post

Observability

Logs, traces, monitoring, run inspection

Observability tools help teams understand what happened after systems and agents ran.

How FST makes it stronger: FST complements observability by controlling what is allowed before it happens, then producing process evidence that logs and audits can reference afterward.

Read the detailed post

Memory systems

Context, recall, knowledge, long-running work

Memory systems help agents remember facts, preferences, prior messages, and task context.

How FST makes it stronger: FST makes memory process-relevant by maintaining official run state: active profile, satisfied gates, missing evidence, approvals, scopes, routes, and next allowed action.

Read the detailed post

The combined stack

Agents do the work. Workflow tools move it. MCP exposes interfaces. FST controls what counts.

1. TriggerA user, agent, webhook, ticket, chat message, or workflow starts controlled work.
2. ExecuteYour agent framework or workflow tool drafts, gathers, reviews, and coordinates the work.
3. CheckFST evaluates the active profile, current run state, scopes, artifacts, gates, and approvals.
4. AuthorizeFST returns the next valid route, asks for missing evidence, blocks progress, or allows materialization.
5. ProveThe run keeps evidence and replay so the team can understand why the work was trusted.

Use the best tool for the work. Use FST for what counts.

FST is strongest when agents are useful enough to touch real workflows, but the team needs scoped authority, trusted evidence, valid approvals, and preflight before protected effects.

Read detailed postsExplore the control loop