Routecraft

Open source AI automation platform  ·  v0.7.0

Where your AI gets to work.

You bought everyone AI licences. The tools your people built with them live on their laptops, on their own logins, and nobody else can use them. Your competitors ship capabilities in an afternoon. You are still arguing about where to run them. Routecraft is the open source platform your teams build on together. What one team proves, every team can use.

Put your AI to workGitHub
Your systems
as they are, where they are
CRMERPHR and payrollsupport deskknowledge basemail and calendarchatsource control
Every way in
when someone asks, and when nobody does
editorMCPCLIHTTPcronwebhookmaileventsfiles
one per person
harness
Local harness
on your laptop, on your own access
Team harness
always on, on service credentials, every call on record
09:14:02 crm:read ana via agent ok
09:14:05 mail:send ana via agent ok
09:15:40 cron digest ok
capability
query CRM
capability
read invoice
capability
send mail
capability
payroll
tool
query CRM
capability
query CRM
tool
read invoice
capability
read invoice
tool
send mail
capability
send mail

Harnesses that work together are a platform.

IThe problem

Your people are not short of ideas. They are short of a place to run them.

Somewhere in your organisation a script already calls the CRM. Down the hall another team is writing it again, on a session token copied out of a browser. Neither knows the other exists. Both run on somebody's personal login, and both stop the day that person is on holiday.

Seen across these
token pasted into a .envbrowser session scrapedone API key for everyonemodel keys in the configexpires daily, log in againno retries, no rate limita handful of tests, or nonewho called? nobody knowsone author, one laptopnothing another team can install
invoice-chaserPython
one VM, one shared key for everyone
recruitment-agentNode
a laptop, a person's token, off on Fridays
expense-approvalsNode
logs in as you, copies the session cookie
support-repliesPython
its own prompts, its own model key
mcp-tools-repoTypeScript
works on one machine, nobody can install it
payroll-checksPowerShell
run by hand every Friday, on one login
agent-skillsMarkdown
the same skill, written four times
sales-followupsNode
reads the CRM on the rep's own login
the next one
being written this week
the one in your area
not on any list yet
a public model API?a key in the config, nobody can tell if it is set
CRM
ERP
HR and payroll
support desk
knowledge base
mail and calendar
chat
source control

Same problem, a stack per team, a credential per person, and nothing another team can install.

Every team builds its own: a stack per team, a credential per person.
IIThe cost

You are paying for the licences. You are not getting the results.

  • Build it once, and every licence you already pay for can call it.
  • Your AI burns tokens sifting through content instead of solving the problem. A capability answers in one call, the same way every time.
  • Nothing to replace. Your assistants keep working, with hands.
IIIThe platform

One runtime, every way in, the same capabilities everywhere.

Prove it on a laptop this afternoon. Promote it, and from tomorrow every team, editor and agent in the company calls the same one.

  • One place your AI tools run
  • One record of every call
  • Nothing built twice
  • People keep the decisions
Every way inpeople in their editor, any agent over MCP, the CLI, HTTP, and triggers that need nobody
one per person
Local harnesson your laptop, on your own access. Try anything first.
Team harnessalways on, on service credentials, every call on record
promote when provenremote
shared as
packages
capabilities,
skills, agents
asks you when it needs a decisionpersonal credentialsservice credentials
Your systemsas they are, where they are
CRMERPHR and payrollsupport deskknowledge basemail and calendarchatsource control
One runtime, every way in, the same capabilities everywhere.
01

Ship it this afternoon

Build and prove a capability on your own machine, on your own access. No approval, no waiting, nothing to ask for before you try it.

02

Run it for everyone

Promote it and it runs always on, on your own infrastructure, on service credentials the platform holds and no person does. Every call carries who asked. Switch telemetry on and every call is on record.

03

Never build it twice

Capabilities, skills and agents ship as packages. The next team installs yours instead of writing their own.

04

Where your people already work

Your editor, any MCP client, the CLI or plain HTTP. Or on its own, from a schedule, a webhook or an email. When it needs a decision, it asks a person.

IVBounded by design

Hands, not keys.

Agents have deleted production databases trying to be helpful. On Routecraft the agent is given capabilities, not credentials. Each one does one thing, says who may call it, and runs where the agent is not, on credentials the platform holds. The agent reaches exactly the capabilities your team exposes to it, and nothing else.

Keys
agent→
databaseemaildeploypayments

can do everything the credential can do

Hands
agent→
send_company_email
input
policy
identity
intent
look_up_record
input
policy
identity
intent

can press the buttons you built; cannot build new buttons

Keys open everything behind them. Hands only press what you built.
// One of the hands the agent can reach for.
craft()
  .id('publishBrief')
  .input(BriefInput)
  .authorize({ roles: ['editor'] })
  .from(direct())
  .transform(redactPII)
  .to(http({ url: '/feed' }))

The capability is the boundary. What you expose is all the agent can reach.

You review the capability, not the prompt. A prompt is a request. A capability is a decision, and it stays decided.

Give your agent hands
VThe path

Your first capability, before lunch.

One command, and the assistant you already use can call a capability you just wrote. When it earns its place, promote it, and the next team installs it instead of writing their own.

VIFrom the blog

From the field.

Tutorials, postmortems, and small notes from building Routecraft in production.

The full archive
ai-agentsorganizationsmcp

AI agents are still single-player. Your organisation isn't.

The personal-assistant wave gave every employee an agent with their own access, their own memory, and their own chat window. That model cannot serve a team. The unit of organisational AI leverage is the governed capability, not the personal agent.

7 min readby Jaco Botha, Founder, DevOptix
Read the post

Routecraft is the platform. DevOptix is the team that builds and runs it with you.

AI automation built and operated for SMBs and consultancies, by the people who build the platform.

Talk to us