PitchAI Agent Engine

What it does

Done-for-you service

PartlyYou run the software, but the vendor's people are part of delivery: consultants help write task descriptions on the mid tier and its developers build a missing connector within a week.Consultant help with clear task descriptionsagents.pitchai.net ↗

Not established: Writes code

How much it does unattended

Runs autonomously

YesA described task runs on its own schedule and returns a finished result, though sending, submission, signature and sensitive changes can be kept as human checkpoints.Describe your task once. Set the schedule. Watch it run automatically.agents.pitchai.net ↗

Agent permissions

PartlyOutward-facing actions are off unless explicitly granted for that process, so the default boundary is read-and-prepare with named human checkpoints before anything leaves the company.Only when that authority has been explicitly agreed for the process. Sending, submission, signature, purchase and sensitive changes can remain human…agents.pitchai.net ↗

Not established: Multi-agent

Which models it runs on

Claude

YesAnthropic's Claude and Claude Code both appear in the vendor's model and agent inventory, with no tier restriction attached to the list.Anthropic Claude · Claude Codeagents.pitchai.net ↗

GPT

YesOpenAI GPT and Codex are named in the same model inventory, and the edge-privacy page names OpenAI again as an approved external processor.OpenAI GPT · Codexagents.pitchai.net ↗

Open models

PartlyOpen-weight families DeepSeek, Mistral, Qwen, GLM and Llama are listed as usable models, but the vendor never says whether they run in its own cloud or where.DeepSeek DeepSeek-V3 · R1agents.pitchai.net ↗

Model choice

PartlyThe Engine, not the buyer, chooses a model per task; the vendor's claim is portability of the task definition across models rather than user-level model selection.The Agent Engine is not tied to one AI company or one agent. For every task it picks the model that suits it bestagents.pitchai.net ↗

Where you use it

In a browser

YesWork is described and scheduled in a hosted application; the trial link opens dispatch.pitchai.net, which I loaded and which redirects to a Microsoft sign-in page before the app.Describe your task once. Set the schedule. Watch it run automatically.agents.pitchai.net ↗

Not established: In your editor, On the command line, In your pipeline

Whose machine it runs on

Self-hosted

PartlyOnly the quoted Enterprise tier offers infrastructure you operate yourself; the three published plans are hosted by the vendor with no self-hosting option.Dedicated nodes, or a dedicated node you operateagents.pitchai.net ↗

Not established: Open source

What it costs to run

How it meters

YesBilling is per credit consumed by work that finished, with monthly credit allowances on the paid tiers rather than per-seat pricing.Only successful work uses credits. Agree the completed result and its credit cost before the process runs.agents.pitchai.net ↗

Free tier

PartlyYou can start without a call or a setup fee using a small trial credit, but the grant is €5 and the entry plan then charges for each completed job.€5 free trial creditagents.pitchai.net ↗

Not established: API access, MCP server, Bring your own key

Buying it for a team

A company can buy it

YesTwo of the four tiers are explicitly aimed at a company's teams and processes, and Enterprise adds reserved capacity and a service-level agreement.Bring your team’s processes together with a PitchAI consultant.agents.pitchai.net ↗

Seat model

NoNo tier is priced or limited per user: all three published plans quote a monthly fee against a monthly credit allowance, with no seat count, floor or ceiling anywhere in the table.€400per month 600 credits per monthagents.pitchai.net ↗

Pooled budget

PartlyEach paid plan carries one monthly credit pool for all of the account's processes, but the vendor never describes per-user allocation or limits within that pool.4,000 credits per monthagents.pitchai.net ↗

Admin controls

PartlyConstraints are agreed per process as human checkpoints and a completion rule rather than enforced through any admin console or policy setting the vendor describes.Agree a clear beginning, end and completion rule before work starts.agents.pitchai.net ↗

Audit log

PartlyEach run returns its result with sources and completion evidence for review, which is per-run traceability; no account-level log of administrative actions is described.Returned with the work Result, source and completion evidenceagents.pitchai.net ↗

Not established: Single sign-on

What happens to your code

Opt out of training

YesThe vendor states flatly that no model is trained on customer data, and its edge boundary is designed to keep identifying values from reaching a model at all.Never trained on your data No AI model is ever trained on your data.agents.pitchai.net ↗

Data residency

YesProcessing is stated to take place inside the European Economic Area, though no specific country, region or sub-processor list is published.Processing in the EEA Your data is processed within the European Economic Area.agents.pitchai.net ↗

Certifications

PartlyGDPR compliance is asserted for processing, but no SOC 2, ISO 27001 or other audited certification is claimed on either published page.GDPR compliant Processing complies with the GDPR, Europe's privacy law.agents.pitchai.net ↗

Not established: Getting out