Skip to content
iWorldAfric
+254 794 545 474Talk to an engineer
Menu

On the homepage

+254 794 545 474support@iworldafric.com

Locus

AI with the context to move delivery forward.

An AI-assisted delivery workspace connecting repository knowledge, issue preparation, implementation handoffs, design changes and QA evidence to GitHub Projects.

Request a Locus inviteExplore the workflow

Available by invitation.

System overview
  • Repository-grounded AI
  • Design-to-test intelligence
  • Agent & MCP workflows
  • Evidence-led QA

Carry the context from the first question to the verified result.

Locus brings AI into the points where delivery teams need context: understanding a repository, preparing an issue, investigating a fix and identifying tests affected by a design change. That work stays connected to explicit decisions, implementation ownership, QA evidence and the release history.

Understand, prepare, implement and verify.

  1. 01

    Understand the work

    Ask questions with cited repository context. Prepare issues with classification, duplicate review and repository-specific routing.

  2. 02

    Prepare the handoff

    Build a code-backed triage brief, capture decisions and generate an editable implementation prompt with task-specific model routes.

  3. 03

    Verify the change

    Review design-driven test proposals, run the test cases and keep evidence, bugs, retests and results beside the work.

Repository context turns an AI answer into something you can inspect.

The repository Assistant grounds answers in authorised code and linked issue context, with source citations. AI-assisted intake can suggest issue fields, surface duplicates and prepare a destination plan for work spanning repositories.

  • Answers with inspected source references
  • Repository, type, priority, size and label suggestions
  • Reviewable repository moves and linked-issue plans

Repository destination moves and companion-issue creation have explicit confirmation steps.

System overview

A design change can reach the test cases it affects.

Locus processes design-release changes into summaries, release notes and proposed test-case changes. Teams compare the current and proposed definitions with supporting file evidence before an approved revision moves into workbook writeback.

  • Diff-based summaries and affected-test screening
  • Create, update and retire proposals with review history
  • Approved workbook changes with conflict checks and readback

Definition approval and test execution remain separate, recorded steps.

System overview

Give people and AI tools the same delivery context.

Code-backed triage brings proposed causes, scope, files and tests into an editable brief. Implementation prompts, ownership claims and permission-aware MCP tools carry that context into the developer’s workflow.

  • Editable issue and iteration implementation prompts
  • Task-specific provider, model and reasoning settings
  • Session-bound ownership, recorded handoffs and MCP access

Suggested analysis, authorised actions and verified results remain distinct.

System overview

Built around the work.

AI preparation and investigation

Classify an issue, inspect possible duplicates, review repository destinations and assemble a cited triage brief. Open questions and proposed findings remain visible for the team to resolve.

Planning and model control

Configure providers, models and reasoning by task. Iteration placement, backlog grouping and plan/build/review model routing work within the configured rules and authorised scope.

Design-release intelligence

Connect versioned design releases to diff summaries, release notes, affected-test screening and reviewable create/update/retire proposals. Preserve the reviewed revision and subsequent writeback state.

Permission-aware AI tooling

MCP tools expose repository context, issues, evidence, test cases, decisions, prompts and ownership workflows to authorised AI clients. Project grants, roles and operation-specific checks remain in effect.

QA with a distinct verdict

A closed GitHub issue can still await verification. Connected test workbooks, assignments, build/environment details, evidence, linked bugs and retest history support the actual QA decision.

Engineering oversight

Repository security scans, dependency inventories and scan reports sit alongside delivery and reopen analytics. Owners can review team results and activity; these records support inspection rather than promise an automatic release approval.

One workspace, several kinds of responsibility.

Engineering and product teams

Bring repository questions, issue preparation, planning, design releases and implementation context into one connected workflow.

QA leads and testers

Keep case definitions, execution evidence, failed attempts, bug links and retests connected to the change under review.

Platform owners

Set AI-task policies, provider and model choices, project access and ownership boundaries for people and connected AI tools.

Explore Locus with your delivery team.

Locus is available by invitation. Bring your GitHub Projects, design and testing workflow, and the AI tools your team uses. We can walk through the context, controls and handoffs together.

Request a Locus inviteDiscuss your delivery workflow

Tell us what you’re building or improving.

Call or email, and a senior engineer will reply in under 6 hours.

Project enquiries