Pydantic AI

Not in the preview yet. This feature is built in the repository but is off in the v2.2.6 preview (test mode). Do not treat this page as available on a preview core. See ../preview.md and the release notes.

dynamoprotocol-pydantic-ai adds Dynamo to Pydantic AI. Every agent in a run gets its own capped slice of one budget. A hand-off or an agent called as a tool opens a child slice, which can never exceed what its parent has left. A model request that would go over its slice is refused before the model is called.

The plugin is not published. It installs from a clone of the repository.

Install and try it

These are the commands from the plugin README. They boot a local core in Control Mode, build a venv from the hash-checked locks/test.txt and run the offline containment demo on a scripted mock model: no provider key, no model called. You need Node 22.12 or later, pnpm 9, curl, openssl and Python 3.11 (set PYTHON=python3.11 if python3 is another version). Set DYNAMO_REPO to your clone.

cd "$DYNAMO_REPO"
pwd
pnpm install --frozen-lockfile
pnpm --filter @dynamo/core-host... -r build
export DYNAMO_OWNER_TOKEN="dp-local-owner-$(openssl rand -hex 16)"
export DYNAMO_CORE_URL="http://127.0.0.1:${DYNAMO_CORE_PORT:-8790}"
[ -s .dynamo-local.key ] || (umask 077 && openssl rand -hex 32 > .dynamo-local.key)
export CORE_KEY_ENCRYPTION_KEY="$(cat .dynamo-local.key)"
OWNER_API_TOKEN="$DYNAMO_OWNER_TOKEN" PLUGIN_WIRE_TOKEN="dp-local-wire-$(openssl rand -hex 16)" \
  OWNER_API_PORT="${DYNAMO_CORE_PORT:-8790}" CORE_DB="$PWD/.dynamo-local/core" CORE_LOG=off \
  CORE_RATE_LIMIT_RPS=200 CORE_RATE_LIMIT_BURST=400 \
  node deploy/image/entry.mjs > .dynamo-local.log 2>&1 &
CORE_PID=$!
trap 'kill "$CORE_PID"' EXIT
"${PYTHON:-python3}" -m venv .venv-pydantic-ai
PIP_ONLY_BINARY=":all:" .venv-pydantic-ai/bin/pip install --quiet --require-hashes -r packages/plugin-pydantic-ai/locks/test.txt
.venv-pydantic-ai/bin/pip install --quiet --no-deps ./packages/sdk-python ./packages/plugin-pydantic-ai
for _ in $(seq 150); do curl -fsS "$DYNAMO_CORE_URL/healthz" > /dev/null 2>&1 && break; sleep 0.2; done
.venv-pydantic-ai/bin/python packages/plugin-pydantic-ai/examples/containment/run.py
kill "$CORE_PID"; trap - EXIT

Delete .venv-pydantic-ai when you are done.

Wrap an agent run in a budget

This is packages/plugin-pydantic-ai/examples/app/with_dynamo.py: three agents that hand off through output functions, with Dynamo added. make test-plugin-pydantic-ai runs it against a local core (tests/test_examples.py).

# SPDX-License-Identifier: Apache-2.0
"""A small multi-agent app on Pydantic AI — with Dynamo.

Runs offline: the scripted mock model stands in for a real one.
"""

import asyncio

from pydantic_ai import Agent, ToolOutput

from dynamoprotocol_pydantic_ai import AgentConfig, DynamoPydanticAI, usd
from dynamoprotocol.testing import mock_prices
from dynamoprotocol_pydantic_ai.testing import ScriptedModel, Text, ToolCall, Upstream

upstream = Upstream()
SCRIPT = {"orchestrator": [ToolCall("transfer_to_researcher", {"request": "find sources"})], "researcher": [Text("three sources found")], "writer": [Text("draft written")]}


def model(name):
    m = ScriptedModel("mock/scripted-1", upstream, agent=name)
    m.queue.extend(SCRIPT[name])
    return m


def handoff(name):
    async def transfer(request: str) -> str:
        return (await agents[name].run(request)).output

    return ToolOutput(transfer, name=f"transfer_to_{name}")


agents = {}


async def main():
    dyn = await DynamoPydanticAI.connect(
        prices=mock_prices(), cap=usd("5"), mandate={"categories": ["research", "writing", "review"], "maxDepth": 3},
        agents={"orchestrator": AgentConfig(usd("5"), "review"), "researcher": AgentConfig(usd("1"), "research"), "writer": AgentConfig(usd("1"), "writing")})
    for name in SCRIPT:
        outputs = [str, *(dyn.handoff(o) for o in SCRIPT if o != name)]
        agents[name] = Agent(model(name), name=name, output_type=outputs, model_settings={"max_tokens": 256}, capabilities=[dyn.capability(name)])
    result = await dyn.run(agents["orchestrator"], "Write a brief on grid-scale storage")
    print(result.output)
    await dyn.close()


asyncio.run(main())

agent.

calling agent, then runs key's agent.

default_completion_tokens to connect; a call with neither is refused (unbounded). The worst case of each call is reserved before it is made.

DYNAMO_CORE_URL and DYNAMO_OWNER_TOKEN point it at your core. With a real model, pass your price table instead of mock_prices(); an unknown model is refused.

What a refusal looks like

A refusal raises DynamoCapExceeded, DynamoMandateRefused, DynamoRevoked, DynamoUnpricedModel, DynamoFrozen, DynamoStopped, DynamoApprovalPending, DynamoApprovalDenied or DynamoAmendmentRefused, all subclasses of DynamoPydanticAIError (a Pydantic AI AgentRunError). It is raised before handler() calls the model.

Funding modes

always a Control Mode budget: every check runs and no money moves. The quickstart and every test of this plugin use one.

budget by id (budget_id=). The plugin does not refuse it, but no test runs the plugin on one. In those modes the core requires signed hops by default, and a child slice opened without its delegator's signature is refused hop_unsigned. Pass hops= (a LayerHops from dynamoprotocol.agents: the delegator's key and each payee's own signed offer) so the plugin signs every hop it opens.

operator allows it, and real-money settlement is not released.

Known limits

Read packages/plugin-pydantic-ai/KNOWN-LIMITS.md in the repository before you rely on the plugin.