Try Professional MCP server
Commit Nov 11, 2026Organizes AI agents into a hierarchical company structure with specialized roles, shared memory, and human-in-the-loop approvals.
Also known as goranai/try-professional
- Stars
- 2
- Tools
- —
- Downloads, 30 days
- —
- Latest version
- —
- Last commit
- Nov 11, 2026
- License
- MIT
What is Try Professional MCP?
Try Professional organizes AI agents into a hierarchical company structure with specialized roles, shared memory, and human-in-the-loop approvals.
Summary generated from the project's README
About Try Professional MCP server
Organizing multiple AI agents into a hierarchical company structure is what this server enables for developers who need more than a single assistant. Instead of one agent, you interact with a CEO who manages specialized departments, managers, and workers, each possessing its own memory and specific skills. This structure allows for complex workflows where roles like QA or legal counsel can review work before it is finalized.
You can deploy this system through Claude Code, the terminal using the trypro engine, or via Telegram. It supports a multi-model approach where different roles use different providers; for example, researchers might use Gemini in Antigravity while engineers use GPT in Codex. The setup requires a GitHub token for skills and can use OpenAI-compatible keys for local execution. While it automates much of the process, it enforces human-in-the-loop approvals for sensitive actions like creating pull requests or changing global git settings, ensuring you maintain control over the final output.
What you can do with it
- 1Designing a business structureAsk the system to architect a complete organizational chart with specific departments and roles based on a single business idea.
- 2Managing automated workflows with oversightSchedule recurring tasks that pause for manual approval before performing external actions like pushing code or sending emails.
- 3Coordinating diverse AI modelsAssign different LLMs to specific roles within a single company to optimize for cost or capability.
Written from the project's README. Every use case is checked against a quote from it.
Automated checks
Ran Nov 15, 2026Full report →Not verified yet. A server needs a score of 80, no serious finding, and enough data for us to check. How checks work · What to fix for more points
- Source repository linked and reachablegoranai/try-professional is public and reachablegithub.com/goranai/try-professional ↗Pass
- Package names this repository as its sourceNot checked
- Build provenance (Sigstore attestation)Not checked
- Open source license (OSI)MITspdx.org/licenses/MIT.html ↗Pass
- Commit or release in the last yearLast commit or release 2026-11-11github.com/goranai/try-professional ↗Pass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNot checked
- No critical advisory in the package or its dependenciesNot checked
- No hidden instructions to the model0 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNot checked
- Remote endpoint answers the MCP handshakeNot checked
README
View on GitHubTry Professional
An agent, but a company. You talk to the CEO. Managers split the work, workers do it with real tools, QA and managers check it, and the CEO answers you. Every role is a full agent, with its own memory and skills.
Preview: proven on Claude only. It works, and it has rough edges. What was tested.
You say what you want, and a company is designed for it: departments, roles, and skills from GitHub that are checked before they're used. You approve it, and then you talk to it, from the terminal, from Claude Code, or from Telegram. It works when you ask, and it remembers: every role, from the CEO down, keeps its own memory, and lessons from each job carry into the next.
Where its roles work:
- Claude Code, on your Claude plan. This is the most tested.
- Codex or Antigravity, on their own plans or keys, with each role's step one non-interactive run of it.
- Its own engine, with any OpenAI-compatible key, such as DeepSeek, Grok, or OpenAI:
trypro chat. Cheap models finish less often today.
They can be mixed, role by role. Say a company coordinates in Claude Code, its researchers look for bugs with Gemini in Antigravity, and its engineers fix them with GPT in Codex. One company, one memory, and its reviewers can be a different model from whoever did the work. In trypro.yaml:
runtime:
default: claude-code # every role, unless named below (with none: trypro's own engine from a terminal, subagents in /company)
roles:
research: antigravity # a department
engineering/backend: codex # one role
models:
antigravity: gemini-3.8-flash-high
From Claude Code, /company hands those roles' steps to their agent, and every other role is a subagent. From a terminal, trypro run does. Codex uses the company's own endpoint and key from .env when there is one, or its own login; Antigravity and Claude Code use their own. Antigravity runs a role's commands without asking unless TRYPRO_ANTIGRAVITY_BYPASS=0.
Whatever the agent, gh and git run what only reads, and git's work on local files anywhere on the machine, such as a commit or a clone. Anything else, such as a pull request, an issue, a release, a push or a change to git's global settings, waits for you in trypro held. Variables named like keys, tokens, secrets or passwords don't reach a role's commands, apart from a GitHub token. On Windows, Codex's and Antigravity's own sandboxes didn't hold in our tests, so a role there can reach the rest of the machine.
| Talk to it | How |
|---|---|
| Claude Code | Open the company's folder in Claude Code, and say what you want |
| The terminal | trypro chat |
| Telegram | trypro telegram, with a bot token in .env |
Who it's for
Anyone who wants a team of agents, organized like a professional company: you give the CEO a job, a department's manager hands it to a worker, the worker does it with real tools (the shell, git and gh, files, and the web), QA and the manager check it, and the CEO reports back to you. Each role remembers its own work, uses its own skills, and leaves an audit trail.
For an everyday one-off task, use a single agent. It's cheaper and faster. A hierarchy spends tokens on reviews, and that's only worth it when a mistake would reach someone outside.
Demo
A real run on 2026-11-08, on Sonnet through Claude Code. Waiting is cut, and each cut says how long it was. Long output is shortened, and marked. Claude Code ran non-interactively, and a script answered the real review screen, standing in for the person who approves. How it was recorded.
What a company makes
These are real, from the journey. This catering company was designed from one line, "a catering business that delivers office lunches", and approved:
Catering
├── Executive: CEO (head) · Chief of Staff
├── Operations: Ops Lead (head) · Support · Menu Planner · Delivery Scheduler
├── Finance: Finance Lead (head)
├── Marketing: Marketing Lead (head) · Content Writer
├── Sales: Sales Lead (head) · Proposal Writer
└── Legal: Counsel (head) · Compliance Reviewer
Asked for "a weekly lunch menu and a price list for a new office client with 30 people", its menu planner wrote this. The ops lead approved it for QA, and Counsel for Legal:
Mon - Lemon-herb grilled chicken, rice pilaf, green salad. Allergens: none of the 14 major allergens in the base
recipe (check stock cube/dressing). Halal: yes if halal-certified chicken is sourced. Not vegetarian.
Tue - Vegetable lasagna with side salad. Vegetarian. Allergens: gluten, milk, egg, celery.
Wed - Chickpea & spinach coconut curry with basmati rice. Vegan, vegetarian, halal, gluten-free.
...
PRICE LIST (USD, excl. tax and delivery)
Item | Per person | Total for 30
Mon Chicken pilaf | 11.50 | 345.00
...
WEEK TOTAL | 55.50 | 1,665.00
It says what it doesn't know: prices are placeholders, and the allergen notes are for a kitchen to confirm. The whole menu, and its index.
It works in the command's language. For a banana-chip business in Indonesia, asked about shipping to Surabaya, a reply that names no rate it doesn't have:
Halo Kak, terima kasih sudah menghubungi kami! Kami kirim ke Surabaya. Besar ongkir tergantung berat pesanan
dan kurir yang dipilih. ...
The whole reply, and a two-week Instagram plan in Indonesian.
Quickstart
For Claude Code. You need Python 3.10 or newer, and Claude Code signed in. It uses your own Claude plan, with no API key, and each role runs as its own subagent.
pip install "try-professional[mcp]"
mkdir my-company && cd my-company
trypro start "a weekly newsletter for my yoga studio's members"
Say what you want made, in your own words. trypro start checks your setup, has Claude Code design a company that can do it, and shows you that company to approve: type a, then approve. A person has to, and no model can. Then it opens Claude Code on your request. When Claude Code asks whether you trust this folder, say yes. Results are files, in deliverables/<run-id>/. Start with index.md.
After that, start Claude Code in this folder, and say what you want. You don't need /company: the folder's front desk hands every request to the company, and the company remembers its earlier work. A quick one-off whose answer stays in the chat, such as a fact to look up, the front desk does itself, and says so.
Running it without the chat, with claude -p, as in a script or CI: open Claude Code in the folder once first, and accept the trust prompt. Until a folder is trusted, Claude Code leaves out each role's own tools, and roles stop with NO_ROLE_TOOLS.
Step by step
The same, one step at a time.
1. Install it, and make a folder for your company.
pip install "try-professional[mcp]"
mkdir my-company && cd my-company
2. Set it up for Claude Code. This checks Python, Claude Code, and your GitHub token, says how to fix anything missing, and adds Try Professional to Claude Code for this folder.
trypro setup claude-code
If it says GITHUB_TOKEN isn't set, set it before you start Claude Code. Without one a company gets no skills. With the GitHub CLI, that's export GITHUB_TOKEN=$(gh auth token).
3. Describe your business. Claude Code designs a company and chooses skills, and then gives you a proposal id.
claude "Use the trypro tools to design a company for a catering business that delivers office lunches."
4. Read the proposal, and approve it. A person has to do this, in a terminal. No model can. Approving also creates the company's roles in Claude Code. The first approval downloads a local embedding model, about 0.22 GB. Your text is embedded on your machine and isn't sent anywhere.
trypro init --review <proposal-id>
5. Start Claude Code again in this folder, and give the company a command. When it asks whether you trust this folder, say yes. Otherwise each role's server can't start, and /company stops rather than do a role's work itself.
/company Write a weekly lunch menu and a price list for a new office client with 30 people
Your session only coordinates. Results are files, in deliverables/<run-id>/. Start with index.md.
Once a company exists
/company <a command> Give it more work. It remembers earlier work
/company continue <run-id> Carry on with one that stopped, in a new chat
trypro doctor Check your setup, and how to fix what's wrong
trypro status Memory size, and tokens by month and command
trypro audit <run-id> Verify a run's audit log
trypro cancel <run-id> Stop it. Its records stay
Everything it needs you for, it asks in the chat. When a command stops for you, the CEO proposes a new role, or it proposes something the whole company should remember, Claude Code shows you, you answer in your own words, and it asks you to confirm before your answer is recorded. When you say something should always or never happen from now on, it offers to make that a standing rule: every role sees it on every step, so you never repeat it in a command. It asks every time, whatever permission mode you're in, and no role can answer for you. If you post something yourself through the chat, it's recorded for the role whose work it was. After an upgrade, its files in .claude/ catch up by themselves the next time Claude Code starts, keeping the model you gave each role. In a terminal, trypro decide, trypro org review, trypro memory review, and trypro rules do the same.
When it works on its own
Whatever the work is, what a company sends out, such as a pull request, a post, a quote to a client, or an email, is tracked with who it went to. What happens to it, with the reason, goes back to every role that worked on it, and the next job for the same client, repository, or site starts from it. A reply becomes a follow-up command. Commands can run on a timer, and while nobody is watching, anything that changes something outside waits for your approval.
trypro schedule add daily "<what to do each time>" --every 1d Run a command on a timer
trypro schedule install --apply Let your system's scheduler run the rounds
trypro schedule run Or run one round now: --watch keeps going
trypro held What waits for your approval
trypro held approve <number> Run it. Or: trypro held deny <number>
trypro outcomes What it put out, and what came of it
trypro outcomes check Check GitHub links now, with gh
trypro outcomes set <number> rejected --note "<what was said>" Say what happened, where no check can
trypro outcomes add "<link or reference>" --run <id> --role <role> --target "<client>"
- Tracked: a link printed by an outside command you approved, such as the pull request
gh pr createopened. In Claude Code, a role gives a link or any reference, and who it went to, tooutcome_track. A target is a repository, a client, a site, or anything a role names. - Verdicts, for any work: accepted, rejected, duplicate (they already had it), by design (it's how they mean it to be), low impact (right, and not worth acting on), or ignored. You give one from the terminal or chat, with what was said:
/outcome 3 rejected "went with a cheaper quote". - Checked by code, where it can be: GitHub issues and pull requests, through
gh, with no model, is the checker that ships. It reads GitHub's own record, labels, and a maintainer's last words, links a duplicate to the issue it duplicates, and calls 30 days without a reply ignored. - Learned from: each role that worked on it gets the lesson, with the reason, in its own memory, and each department head decides whether the department keeps it. When a command names an earlier command's run, its roles are credited too. Roles that plan, assign, do, or check work see the record, with whatever their work names first: that client's, repository's, or site's history, and what it doesn't want.
- Checked before it goes out: whoever decides on something outside, you at the prompt, or later in
trypro held, and the reviewers, sees what the company already sent to the same target, and what came of it. Searches can plug in beside that; one that ships looks for issues a new GitHub issue may duplicate. - Followed up: a new reply from someone else queues a command to read it, and answer only if an answer is needed. That answer needs your approval, like anything else.
- From chat:
/held,/approve 3,/deny 3,/outcomes,/outcome, and/scheduleswork intrypro chatand on Telegram. With a Telegram bot set up, each round's summary is sent there. - Scheduled rounds use the company's own engine, with the model
trypro.yamlnames. Claude Code isn't run unattended.
When it needs someone new
A company starts with the roles you approved. When its work shows one is missing, the CEO proposes it, and you decide: in Claude Code, it asks you at the end of the command. Nothing changes until you do. From a terminal, trypro org proposals lists them, and trypro org review decides them.
- When it proposes: at intake, when part of a command fits no role, and at the end of any command that was sent back, escalated, or handed between departments, and of every large one. It sees every department, and the record: work sent back and why, departments left with no work across recent commands, and work put out by a person by hand. It proposes only on that evidence, never for one-off work, and never a role that exists. An idle department isn't a missing role: if it should have worked, that's a lesson about planning instead.
- What it can propose: a role in the department whose work it is, or a new department with its role as the head, when no department fits. A new department's work passes the same review as the others'.
- When you add one, it's checked like any role, and undone if it would break the company. In Claude Code, the roles are rewritten, and each keeps the model you gave it. A rejected one is never proposed again.
- A command that names another department's role, such as "confirmed by the validator", is planned for both departments, even when the CEO sized it for one. A real company's CEO sized six hunts that way, and its Validator never ran.
Build on Try Professional
A company can be a part of your own agent. The Python API is stable across 0.x minor versions; the MCP tools and the adapter interface are experimental. What's stable, the Claude Code path, custom roles, and extending it.
from try_professional import Company
company = Company.create("acme", name="Acme") # the default company: no model, no cost
company.add_role(
"marketing",
"newsletter-editor",
title="Newsletter Editor",
mission="Write the weekly customer newsletter.",
skill_needs=["write email newsletters"],
)
company.setup_claude_code() # then, in Claude Code: /company <goal>
examples/claude-code-agent has a delegate() tool your agent can call, to hand a goal to the company through Claude Code and read back what it made. Approving a company the AI designed, and answering a run's questions, stay with a person.
Multi-agent orchestration for Claude Code
Each piece of work has one owner, gets reviewed, and leaves an audit log you can verify. Each role runs as its own subagent, which sees only its own assignment, memory, and skills. The rules that matter are enforced by code, not by asking a model nicely.
- A company from one description, adapted from five industry templates, in any language. It proposes, and a person approves.
- Skills from GitHub, vetted first.
skill-customspins each one to a commit and checks its license and content, and a model judges which ones fit each role. It can also be used on its own. - A permanent company. Roles have their own memory, departments and the company have shared memory, and later commands get the finished work of earlier ones.
- Every role is a full agent. In Claude Code, each role works with the shell,
gitandgh, files, and the web, and Claude Code asks you before commands unless you allow them.trypro agents claude-code --sandboxedtakes those tools away, for a company that should only write documents. - Reviews that can't be skipped. QA reviews everything. A Legal department is optional, and off unless the business needs one: then its review is triggered by what the work is, whatever a model declares.
- Rules held by code. Work must report on every item of its brief. Reviews can't approve with blockers. Unmet criteria and unverified claims reach you even if a final report leaves them out. There's no token limit unless you set one: a company works until the job is done. It does stop, and say why, when a role keeps going in circles: 24 model calls with nothing moving, or 3 hand-ins refused in a row. To cap spending, set
max_tokens_per_small_commandandmax_tokens_per_rununderpipeline.budgetinorg/org.yaml, and a command stops at them and asks you. Tests play a misbehaving model against these. - No group chat. Specialists keep their own job and memory, and you talk only to the CEO. Roles never message each other: work moves through the hierarchy as hand-ins, so there's nothing to loop on. The CEO uses the fewest roles that can do it well, a head asks another department only for what its own roles can't do, and says why, work can't be handed back, and independent work runs at the same time. Every run reports its minutes, calls, tokens, and extra rounds, so you can see what it cost.
- Real separation over MCP. Each role's server shows it only its own assignment, memory, and skills, and the session's server can't do a role's work.
- It learns from what happened. Whatever the work, what the company sends out is tracked with who it went to, and what came of it, with the reason, goes back to every role that worked on it. The next job for that client, repository, or site starts from it. How
How it works
/company <your command>
|
v
CEO: turns it into criteria someone can check
|
v
Department: does the work <-------- sent back, saying what's wrong
| ^
v |
QA reviews it, and then Legal, if the company has one and the work is public or touches licenses or personal data
|
v
CEO: the final report, in your language
|
v
deliverables/<run-id>/index.md, and an audit log
- Design. Claude Code calls the
tryprotools to adapt one of five industry templates to your business, and to match skills to roles.skill-customspins each skill to a commit, and checks its license and content. - Approve. You read the proposal and approve it in a terminal. No model can do this.
- Roles. Approving, in a folder
trypro setup claude-codeset up, writes one subagent per role. Each has Claude Code's own tools, and a server bound to that role, which gives it only its own assignment, memory, and skills.trypro agents claude-codewrites them again after a company changes. - Commands. Your session only coordinates. It can see who's next, and it can't read an assignment or submit anything. A small command goes straight to one role, and a large one is planned by department heads.
- Memory. Roles keep their own memory, and later commands get the finished work of earlier ones.
What was tested, and how
| Model | How | Result |
|---|---|---|
claude-sonnet-5-5 for every role, and claude-haiku-5-5 for matching skills |
Claude Code 2.1.319 on Windows 11, through MCP, each role as its own process with only that role's server | 9 of 9 commands finished, on 3 businesses. All 108 artifacts were recorded as coming from the right role's own server. Details |
claude-sonnet-5-5 for every role |
/company in one Claude Code session, each role as a subagent, on a business kept out of all training |
One prompt to a company, a command, and a later related command that followed the first. All 23 artifacts came from the right role's own server. It passed on its sixth run, and the fifth stopped at the token limit. Every run |
claude-haiku-5-5 for every role, and for the session |
0.6.3: /company in Claude Code, seven everyday jobs, three times each |
21 of 21 finished, from 1.3 minutes and $0.014 for a customer reply to 5.2 minutes and $0.11 for content that needs research, at list price. Details |
claude-haiku-5-5 for every role, and for the session |
0.1.x: /company in Claude Code, on the acceptance journey, twice |
0 of 6 commands finished. Each stopped for a person after its rework limit, or because earlier work wasn't approved. Isolation held: 90 artifacts, none claimed. Its reviewers caught 22 of 22 seeded flaws, and sent back 2 of 4 clean documents. Details |
deepseek-v4.1-flash-1017 |
API runner | check-model: usable with caveats. The same small command finished once in three runs, and the other two blocked. Why |
step-5-preview |
API runner, 0.6.0 | Real jobs, with the shell and the web: 4 of 5 finished, including one using gh and one in Indonesian. Details |
Those are the only models tested. The test suite runs on a scripted model, on Linux, macOS, and Windows.
This doesn't show that a cheap model does good work. Quality on a cheap model is for v0.2.
Try Professional MCP questions
How do I use Try Professional with Claude Code?
Install the package, run 'trypro setup claude-code' to configure your environment, and then use the '/company' command within Claude Code to assign tasks.
Can I use different AI models for different roles?
Yes, you can mix models role by role in your configuration file, such as using Gemini for research and GPT for engineering.
Does the system require manual approval for actions?
Yes, certain actions like pull requests or changes to global git settings are held for your approval to ensure safety.
How does the agentic memory work?
Every role maintains its own individual memory, allowing lessons and context from previous jobs to carry over to future tasks.