Kortix vs Paperclip: Sessions or Shared Work Records?

6 min read Comparisons

Compare Kortix and Paperclip on work records, approvals and recurring execution, including experimental features and hosting limits.

Should Kortix or Paperclip hold the work when an agent needs to hand a job to a person? Kortix centers agent sessions and repository configuration, while Paperclip publishes task records that people and agents can both own. That distinction matters when a correction needs an owner and a durable discussion.

A session can preserve what an agent was doing, but your team also needs to know who is responsible next. Compare the record that survives a handoff before comparing how many tools an agent can call. Both products publish approvals and recurring execution, so those capabilities alone do not decide this choice.

Kortix starts with shared agent configuration

Kortix puts agents, skills, connectors and memory into repository files. Its guide describes My Kortix delegating work while you continue the conversation, along with editable Markdown memory in Drive. You can inspect what it remembers and ask it to forget information, which matters when a shared agent retains an outdated instruction.

Its experimental agent teams add a lead, teammates and a shared task list. These teams are off by default and last for the lead's session, so distinguish their coordination from a saved team that persists across jobs. Numbered tasks can have owners, dependencies and status, but the documentation does not establish separate work records that people and agents both own with their own discussion and history.

Paperclip gives the handoff a task thread

Paperclip's public overview describes tasks assigned to either a person or an agent, with conversations, plans, files, blockers and run history in the thread. A developer can take responsibility for a blocker on the same task an agent was executing. That is a concrete fit for teams whose work changes hands during review.

Paperclip also organizes agents through reporting lines. A CEO agent assigns issues and managers delegate to their reports. Its published goal model includes targets and linked projects, though it does not describe metric values and trends automatically causing a goal lead to review progress.

A permission gate and a review stage answer different questions

Kortix lets you set Allow, Ask and Block rules for connector actions. Review holds the exact proposed call for a person to approve or deny, and changing the call requires fresh approval. For an outgoing message, that gives the reviewer a specific action to judge before execution.

Paperclip also has a Decisions queue for questions, proposals and approvals. Its action-permissions documentation places Allowed, Ask first and Off settings on the connection. The scope differs from a setting attached separately to each agent, so check which agents share the connection you intend to constrain.

For finished work, Paperclip's execution policies define ordered review and approval stages with human or agent participants. Reviewers can request changes. Kortix's experimental Goals instead use a separate verifier session to judge each done criterion, sending work back when any criterion is failed or inconclusive. These mechanisms answer different needs: routing a deliverable through reviewers and checking it against stated completion criteria.

Scheduled agents do not imply scheduled workflows

Kortix supports scheduled triggers, signed webhooks and loops that wake existing sessions with their history. Its developer documentation also describes JavaScript workflows with pipelines and parallel agent calls. However, the workflow tool must start from a person's message, and a trigger or reminder cannot invoke it.

Paperclip's routines create tasks from schedules or signed webhooks. Its execution policies then provide review stages, not general branching workflow steps. For an unattended recurring report, test how the initial event, production work and reviewer handoff connect, because the presence of a scheduler does not establish that entire sequence.

Kortix adds a second unattended boundary: a triggered session waits at most 30 minutes for a person and then stops. Paperclip's routine documentation does not describe a stable root task with separate occurrence histories. Consider whether your operator needs to answer during the run or inspect repeated jobs afterwards.

Hosting and model costs need separate decisions

Kortix supports hosted and self-hosted deployments, plus a desktop connection for local filesystem, shell and screen access. Its pricing page lists Free with 500 monthly credits on the first workspace and Credits starting at $50 a month for 5,000 credits shared by unlimited members. Compute and models bill separately, managed models add 5.5%, and own-key usage adds 5% above $1,000 a month.

Paperclip's homepage describes MIT-licensed self-hosted software and offers a Cloud waitlist. Cloud documentation exists, but the checked public pages do not establish general hosted availability or publish a Cloud price. You arrange AI access, including a supported subscription or provider key, so open-source licensing is only one part of operating it.

Kortix and Paperclip, point by point

The table presents both products' answers to the same questions, including the source and check date for each answer. Read the qualifications alongside each capability, especially where the documented feature is experimental.

Both columns answer the same questions from each company's public pages. Yes means the company publishes the capability the question asks about, Partly means something narrower, No means it says it does not offer it, and Not published means its public pages do not describe it. Every answer links to its source and shows when we last checked it. The oldest check on this page was November 13, 2026.

Kortix and Paperclip compared, with the source and check date of each answer
What you compare Kortix Paperclip
Goals Kortix Yes

Kortix describes agents running a search program that tracks queries, rewrites declining pages and measures results. Its Goals feature assigns an agent to done criteria and scheduled wakes, but is experimental and off by default.

kortix.com kortix.com/docs/guide
Paperclip Partly

Goals can contain measurable targets, link to projects and have sub-goals. The docs do not describe linked metric values and trends that automatically trigger a goal lead's review.

Leads Kortix Yes

An agent team has a lead that splits work, coordinates teammates through a shared task list and combines their answers. Teams are experimental, off by default and exist only for the lead's session rather than as saved teams.

Paperclip Yes

A CEO agent creates and assigns issues, while manager agents delegate to their reports. The docs do not describe distinct goal and project leads that review outcomes and propose gaps for approval.

Tasks Kortix Partly

Agent teams have numbered tasks with owners, status and dependencies, while sessions keep the conversation. Its docs do not establish work records that both people and agents can own with their own discussion or history.

Paperclip Yes

Tasks can be assigned to an agent or a person, and their threads keep conversations, plans, blockers, files and run history.

paperclip.ing/llms.txt
Recurring work Kortix Yes

Triggers start agents on schedules or signed webhooks, and loops wake existing sessions with their history. Connector events are also available experimentally and are off by default.

kortix.com/docs/guide
Paperclip Yes

Agents wake on heartbeats, and routines create tasks on schedules or from signed webhooks. The docs do not describe a stable root task with separate occurrence histories.

Workflows Kortix Yes

JavaScript workflows coordinate agents through parallel calls and pipelines, and saved scripts can run again with arguments. The workflow tool starts only from a person's message, not a trigger or reminder.

kortix.com/docs/developers
Paperclip Yes

Execution policies provide ordered review and approval stages with human or agent participants, and reviewers can request changes. These are review stages rather than general branching workflow steps.

Autonomy Kortix Yes

Allow, Ask and Block rules control connector actions, and Review collects held approvals. A person approves or denies the exact action, and a changed call needs a new approval.

kortix.com/docs/guide
Paperclip Yes

The Decisions queue contains approvals, questions and agent proposals. Connector actions can be Allowed, Ask first or Off per connection, rather than a per-agent rule for every action.

Learned autonomy Kortix Not published

Its public docs describe editable action rules and a reflection agent that improves skills and instructions, but no proposals to change an agent's permissions based on approval history.

Paperclip Partly

People can capture decision examples with notes and export them for a training pipeline. The docs do not describe automatic proposals to change an agent's autonomy based on approval history.

Memory Kortix Yes

Agents keep memory as editable Markdown files in Drive and read it at the start of later tasks. My Kortix saves preferences and corrections, and you can ask what it remembers or tell it to forget something.

kortix.com/docs/guide
Paperclip Partly

Experimental memory connectors let agents save and retrieve context through providers including Mem0 and Zep. Paperclip has no separate memory page or automatic background memory sync, and the checked docs do not establish that users can inspect or correct the saved memories.

Planning Kortix Yes

My Kortix proposes a workflow plan in conversation before running multiple agents. You approve it first, and its card shows phases, agent count and progress.

Paperclip Yes

Experimental Agent Chat can research and draft a plan with you, then create tasks with the agreed plan and what done looks like. Plan mode gives you a plan to review before implementation.

Intake Kortix Yes

Signed webhook triggers route outside events to a chosen agent or existing session. Software Factory can receive issues and failures into a ticket store and give each ticket a worker.

Paperclip Yes

A routine webhook can create a task and assign it to an agent, and a CEO agent can create and assign issues. The docs do not describe the same project-token intake with an optional proposal first.

Connectors Kortix Yes

Kortix publishes a catalog of 3,000+ apps plus MCP, OpenAPI and raw HTTP connections. You can add custom connectors, and signed webhooks start agent work.

kortix.com kortix.com/docs/guide
Paperclip Yes

Paperclip lists service connectors and supports custom MCP servers. Routines can also start work through webhooks.

Models Kortix Partly

You choose a workspace default, pin a model to an agent or select one per session. Ordered fallback models handle failures, but its docs do not describe automatic model selection based on each task or step.

kortix.com/docs/guide
Paperclip Partly

You choose models and runtimes per agent. Experimental pools route accounts for new tasks and pin later runs, rather than automatically choosing a model for each work stage.

Cloud or your own computer Kortix Yes

Kortix offers its hosted cloud and a self-hosted deployment. You can also connect your own computer through the desktop app for filesystem, shell or screen access, with access controlled on that computer.

kortix.com kortix.com/docs/guide
Paperclip Partly

Paperclip supports local and self-hosted deployments, and its docs describe Paperclip Cloud organizations. Its homepage offers a waitlist. General hosted availability is not established.

Your own subscription Kortix Yes

Every plan accepts your own provider keys or a ChatGPT subscription. The guide also documents Claude Pro and Max subscriptions, whose model calls use no Kortix credits.

Paperclip Yes

Paperclip lets agents use your Claude or ChatGPT subscription, or an Anthropic or OpenAI API key. Connect the provider during agent setup and select the model on the agent.

Pricing Kortix Yes

Free includes 500 monthly credits on the first workspace. Credits starts at $50 a month for 5,000 credits shared by unlimited members. Managed models add 5.5%. Own-key usage adds 5% above $1,000 a month. Compute and models bill separately.

kortix.com/pricing
Paperclip Partly

Paperclip's self-hosted software is open source under the MIT license. You arrange AI access. Its homepage offers a waitlist, and the checked Cloud documentation does not publish a price.

Correctness Kortix Yes

Experimental Goals use a separate verifier session that did none of the work. It checks every done criterion as pass, fail or inconclusive, and anything short of all passes sends the goal back to work.

kortix.com/docs/guide
Paperclip Yes

Execution policies route work through review and approval stages, with change requests returned to the executor. A bundled QA skill describes acceptance criteria and a manual validation plan rather than automatic checks on every task requirement.

For software work

Kortix and Paperclip compared, with the source and check date of each answer
What you compare Kortix Paperclip
Code review rules Kortix Partly

Software Factory verifies changes and adds independent review before a merge gate. Its docs do not specify code review against repository rules.

kortix.com/docs/developers
Paperclip Yes

An experimental GitHub review bot accepts your own code-review guidance for pull requests. Its docs do not describe discovering repository rules and screening suggested rules against examples.

Adversarial review Kortix Yes

Software Factory puts an independent review after the worker's verification and feedback, before a merge gate. Its docs do not describe a separate high-risk change classification for that review.

kortix.com/docs/developers
Paperclip Yes

Execution policies can route an issue to an independent reviewer, and low-trust review can contain work on external pull requests. The docs do not describe two independent adversarial reviewers automatically examining high-risk changes together.

Risk rating Kortix Not published

Its developer docs allow risk labels on submitted Review items, but do not describe a rating of a code change before merge.

kortix.com/docs/developers
Paperclip Partly

A bundled GitHub pull request skill instructs an agent to include risk and rollback information in the pull request. It does not describe a revision-specific Low, Moderate or High merge-risk rating.

Pull requests Kortix Yes

Software Factory assigns tickets to workers that verify changes and open pull requests with demos, followed by feedback, independent review and a merge gate. A person merges unless automatic merging is configured.

kortix.com/docs/developers
Paperclip Yes

Agents can create pull requests and push fixes after review. Experimental review bots publish a check on the reviewed commit, which GitHub repository rules can require before merging.

A fact we cannot find or cannot check is left out, and prices can change after the date shown. Task Machine publishes this comparison and is an interested party. Kortix and Paperclip are named only to identify them, and there is no relationship between the companies. Has something changed? Tell us.

Who each one suits

Pick Kortix if repository-managed agents, editable memory and scripted coordination fit how your developers work. Its session-scoped teams and person-started workflow tool are limits to evaluate before assigning an unattended multi-agent process.

Pick Paperclip if people and agents need shared task ownership and ordered review stages in a system you can host. Hosting brings upkeep, and its public Cloud pages leave availability and pricing unresolved.

Where Task Machine fits

Task Machine addresses the handoff through Tasks that keep plans, discussion, decisions and run history together under a person or agent owner. Its Workflow documentation describes branches, human questions, approvals and verifier checks that return weak work with feedback. That connects recurring execution with explicit result checks when a task needs more than a sequence of review stages.

Task Machine is the wrong pick if you need to import Kortix's scripted coordination, because its docs do not describe an importer for Kortix repository configuration or JavaScript workflow scripts. See Switch from Kortix to Task Machine. Compare Task Machine vs Kortix and Task Machine vs Paperclip against the handoff you actually need. As a next step, use the Workflow documentation to map one recurring report's production, verification and human approval steps.