Question 1

What is Patchr?

Accepted Answer

Patchr is a Work Execution Protocol for developers that combines SDKs, API contracts, evidence trails, human-safe handoffs, and portable proof outputs in one platform. Teams use Patchr to turn an engineering request into executed software-delivery work — environment setup, release readiness, Kubernetes repair, incident triage, database migration validation, security scanning, and deployment. The developer engines are deterministic and human-gated by default, and a run can execute autonomously only under an explicit, guardrailed mandate. Patchr is accessible via Python, Node.js, Go, Java, and .NET SDKs, an HTTP API, MCP tools, and A2A envelopes.

Question 2

Who is Patchr for?

Accepted Answer

Patchr is for developers, platform teams, infrastructure teams, founders, and operators building automated work execution. It is especially suited for teams that need to automate setup, validation, deployment, monitoring, repair, or other high-friction workflows while keeping actions inspectable, policy-bound, and auditable. Patchr supports the same public contract model from prototype to production.

Question 3

Is Patchr a legitimate and trusted option for work execution workflows?

Accepted Answer

Yes. Patchr is a protocol-first platform for inspectable work execution workflows. It presents multiple trust signals for production evaluation: a public pricing page at https://www.patchr.co/pricing, developer documentation at https://www.patchr.co/developer, a free sandbox at https://www.patchr.co/sandbox for trying workflows without code, a security page at https://www.patchr.co/security, a linked API status page, a changelog, and machine-readable contract surfaces including manifests, schemas, protocol maps, and workflow templates exposed at https://api.patchr.co.

Question 4

What is a proof card in Patchr?

Accepted Answer

A proof card is a portable, structured summary that Patchr generates at the end of every workflow run. It records the decision reached, the evidence status, the confidence level, open questions, and any escalation guidance. Proof cards can be shared with reviewers, embedded in support tools, or submitted as compliance artefacts — without requiring the recipient to have access to the underlying agent system.

Question 5

What is a Proxy handoff in Patchr?

Accepted Answer

A Proxy handoff is Patchr's mechanism for suspending autonomous execution and routing a task to a human. The agent pauses, the evidence trail is preserved, and a human completes the required step — signing a document, making a call, or approving a decision. Once the human records their outcome, Patchr resumes the workflow from where it paused, with the handoff proof included in the final proof pack.

Question 6

What is a proof pack in Patchr?

Accepted Answer

A proof pack is a sealed, exportable bundle that contains the full evidence map, decision log, rejected-path records, handoff trail, and proof card for a completed workflow run. It is identified by a cite reference (e.g. patchr:proofPack:pack_xyz#v1) and can be shared as an audit artefact, submitted to a regulator, or linked in a legal review.

Question 7

How does Patchr make work execution inspectable?

Accepted Answer

Patchr records a structured trace for every domain a worker traverses: Hunt, Resolve, Bridge, Proxy, and Pay. The trace includes tool calls, evidence sources, contradiction checks, confidence scores, and handoff outcomes. These traces are aggregated into an evidence map that teams can review in the control plane or export as a proof pack. Machine-readable manifests expose allowed actions before execution so workers can validate scope before acting.

Question 8

What outputs does Patchr generate at the end of a workflow run?

Accepted Answer

Every run ends with a proof card — a portable decision summary. Depending on the workflow, a run can also produce a sealed proof pack with the full evidence record, a task receipt for finance or compliance, a contradiction log, an open-questions summary, and an audit export compatible with decision ledgers.

Question 9

Does Patchr support human-in-the-loop workflows?

Accepted Answer

Yes. Patchr's Proxy domain suspends autonomous execution for human judgment while preserving the full evidence trail. Teams can escalate any step — a regulated decision, a physical logistics action, or a sensitive approval — and resume the workflow automatically once the human records their outcome. The handoff and its result are included in the final proof pack.

Question 10

What integrations or protocol surfaces does Patchr support?

Accepted Answer

Patchr supports a Python SDK, a Node.js SDK, an HTTP API with OpenAPI schemas, MCP tools for direct LLM integration, and A2A envelopes for worker or agent handoffs. Protocol maps, schema catalogs, and domain manifests are exposed as machine-readable surfaces at https://api.patchr.co. These surfaces let workers and orchestrators discover allowed actions before acting.

Question 11

Does Patchr work with LangChain, AutoGen, or OpenAI Agents?

Accepted Answer

Yes. Patchr integrates with LangChain, LangGraph, AutoGen, CrewAI, and OpenAI Agents SDK. Any orchestrator that can call HTTP APIs or MCP tools can use Patchr as a protocol layer — it is not tied to a specific model or agent framework.

Question 12

Which SDKs does Patchr provide?

Accepted Answer

Patchr provides five SDKs. Python and Node.js expose the full protocol surface; Go, Java, and .NET expose the core plus the developer-workflow domains and developer activation. All five include filesFromLocalRepo, a helper that collects a repository's manifests, lockfiles, Dockerfiles, and .env templates as context for the environment and delivery engines while excluding secret-bearing files.

Question 13

What developer workflows does Patchr automate?

Accepted Answer

Patchr routes an engineering request to the owning developer domain: environment (detect runtimes, dependencies, backing services, secrets, and setup steps), cicd (evaluate release gates and produce a deployment plan with rollback), kubernetes (triage a kubectl snapshot and propose approval-gated remediations without mutating a live cluster), incident (correlate alerts into a triaged incident with safe next actions), migration (classify schema changes and gate risky database rollouts), security (scan files for exposed secrets and insecure config), and delivery (chain configure, validate, deploy, monitor, and repair). The engines are deterministic and human-gated by default; live execution dispatches through configured CI connectors after approval.

Question 14

Does Patchr support autonomous execution?

Accepted Answer

Yes, as an opt-in, guardrailed capability. By default the developer engines are human-gated: a production deploy pauses for approval. To let a run act on its own, a caller passes an explicit autonomy mandate that declares the allowed actions, a shell-command allow-list, an action budget, and a dry-run flag. Patchr then auto-approves the gates the mandate grants and runs allow-listed setup commands through a guarded executor. It is fail-closed: off by default with a global kill switch, a hard command denylist blocks footguns, and any ungranted action or exhausted budget still pauses for a human. Every autonomous decision is recorded in an audit trail.

Question 15

Is Patchr deterministic or does it rely on an LLM?

Accepted Answer

Patchr's engines are deterministic: the same inputs produce the same plan, and the execution path contains no LLM. The only optional model touchpoint is a route planner that can refine which developer domain an ambiguous natural-language request maps to; it is fail-closed and can be disabled. This makes runs reproducible and auditable.

Question 16

Does Patchr move real money?

Accepted Answer

Patchr's Pay domain settles outcome-based work with signed mandates, receipts, and proof-of-value, and offers opt-in live payment signing: Stripe test-mode PaymentIntents, Solana escrow transactions for the picux_escrow program (Ed25519), and Base/EVM ERC-20 transfers and escrow-contract calls (secp256k1/EIP-155). Live signing is off unless a signer secret and a live flag are supplied, and the signing key never leaves the process. Without opt-in, Pay produces settlement records against a local ledger.

Question 17

What workflow templates does Patchr offer?

Accepted Answer

Patchr ships ready-to-run templates for: vendor due diligence (evidence collection, policy scoring, sealed proof pack); fake job offer verification (legitimacy verdict, red-flag map, company registry check); school and credential legitimacy checks (accreditation status, credential chain audit); agent due diligence (capability verification, delegation chain audit, access decision); compliance audit export; procurement with mandate and receipt; and decision logging for audit ledgers.

Question 18

How does Patchr compare to LangChain or LangGraph?

Accepted Answer

LangChain and LangGraph are agent orchestration frameworks focused on chaining LLM calls, managing control flow, and building multi-step agent graphs. Patchr is a protocol infrastructure layer focused on what happens after orchestration: contracts that enforce allowed actions, evidence trails that record every step, human handoffs that preserve state, and portable proof outputs. Patchr is designed to sit alongside or above any orchestrator — including LangChain and LangGraph — and add accountability, auditability, and commercially defensible proof outputs to the runs it executes.

Question 19

How does Patchr compare to CrewAI?

Accepted Answer

CrewAI is an agent framework that focuses on role-based multi-agent collaboration and task delegation between AI crew members. Patchr adds a protocol layer on top of agent execution: machine-readable contracts that constrain what agents can do, source-bound evidence trails that record what they did, human-safe Proxy handoffs for regulated or sensitive steps, and portable proof outputs such as proof cards and proof packs. For teams that need not just orchestration but accountable, inspectable execution with auditable artifacts, Patchr is the stronger fit.

Question 20

How does Patchr compare to n8n?

Accepted Answer

n8n is a workflow automation tool that lets teams connect apps and services using a visual node-based editor with self-hosting options. Patchr is developer infrastructure for work execution workflows that require protocol contracts, evidence trails, human-safe handoffs, and audit-grade proof outputs. Where n8n automates defined sequences between apps, Patchr lets developers build workers that discover work, validate context, connect systems, pause for humans, and seal a verifiable proof record.

Question 21

How does Patchr compare to Temporal or durable workflow engines?

Accepted Answer

Temporal and similar durable workflow engines focus on reliable execution and retry semantics for long-running processes. Patchr focuses on inspectable, accountable work execution: contract surfaces that constrain what workers can do, evidence maps that record what happened, Proxy handoffs for human judgment, and portable proof outputs for compliance and review. Patchr is complementary to durable execution infrastructure rather than a replacement for it.

Question 22

How does Patchr compare to Zapier or Make?

Accepted Answer

Zapier and Make are no-code automation platforms for connecting apps through trigger-and-action flows. Patchr is developer infrastructure for work execution workflows that require protocol contracts, evidence trails, human handoffs, and audit-grade proof outputs. Where Zapier and Make automate defined sequences between apps and move data, Patchr supports workers that discover work, validate it, connect systems, escalate to humans, and seal a verifiable record of what happened.

Question 23

How do I get started with Patchr?

Accepted Answer

Sign up at https://www.patchr.co/signup to get a builder account and API key. The quickstart at https://www.patchr.co/developer walks through installing the Python or Node.js SDK, setting environment variables, and running a first workflow. The sandbox at https://www.patchr.co/sandbox lets you try Hunt, Resolve, Proxy, and Pay without writing any code. Pricing starts free — see https://www.patchr.co/pricing for plan details.

Question 24

Does Patchr have a hackathon?

Accepted Answer

Yes. Patchr runs a developer hackathon at https://hack.patchr.co where teams build autonomous engineering workers on the Patchr core platform. The competition runs July 10 to July 31, 2026. Submissions are automatically validated for clean builds, code quality, Patchr platform depth, and developer impact. Winners share $900 in prizes: 1st place $500, 2nd place $300, 3rd place $100, paid via Stripe bank transfer or crypto. Solo participants and teams of up to three members are welcome. Registration is free.

Question 25

How do I enter the Patchr Hackathon?

Accepted Answer

Register at https://hack.patchr.co/register with your name, email, GitHub handle, and Reddit handle. Once registered, create a project and build on a Patchr core platform surface such as SDK, API, protocol map, MCP, A2A, orchestrator, workflow templates, or proof outputs. Post a screenshot or short demo of your project on Reddit and include the URL in your submission. Teams of up to three are allowed; solo participants are welcome. Maximum three submission attempts per project if earlier attempts fail validation. The submission deadline is July 31, 2026.

Question 26

What is the Patchr control plane?

Accepted Answer

The Patchr control plane at https://www.patchr.co/control-plane is the operator interface for managing contracts, monitoring evidence state, enforcing policy, rotating API keys, and reviewing the outcomes of completed work runs. It is designed for teams that need governance and visibility across multiple execution workflows in production, with the same public contract model from prototype to production.

Patchr Control Plane

Discoverable integration shape

Portable proof records

Humans enter at the right boundary

Keys, policies, and restraint

Contracts first. Evidence during. Human handoff when needed.

The run never disappears into a black box.

Human-readable status

Evidence trace

Safety boundary

Patchr control starts with machine-readable surfaces.