Turn scattered product knowledge into shared decisions, approved specs and delivery learnings.
Teklens connects your Jira, your docs and your code into one shared memory. Product, engineering and agents work on it through all four phases. You decide what gets approved.
- 01DiscoverFind what matters
- 02DefineDecide what to build
- 03BuildShip with context
- 04OperateLearn what’s next
Operate feeds Discover: what you learn from delivery goes into the next round.
Product Brain: the shared memory for people and agents.
Jira, Confluence, GitHub, GitLab and watched folders, indexed continuously. Plus the decisions and results from your conversations. Every answer and every agent works from the same state.
One search across all sources finds tickets, sprints, pages, documents, branches, pull requests and saved decisions. Code search works separately at symbol level: semantic for concepts, keyword for identifiers and call sites.
Product Brain in detailThe shared memory for people and agents.
- Jira
- Confluence
- GitHub
- GitLab
- Watched folders
- Across all sources: tickets, pages, documents, pull requests, memories.
- In code at symbol level: semantic or keyword.
- Answers cite the source in the index.
Gather what you know before you decide.
Interviews, notes, support cases and ideas sit in different places. Teklens brings them together, makes them searchable and gives every initiative a frame.
- Initiative planning & phase gates
A charter with goal, value, scope and risks per initiative. Artefacts attach to the initiative and only move to the next phase once you approve.
- Watched folders
A local folder of notes, PDFs or Office files is synced continuously and becomes searchable. The watcher runs on macOS, Linux and Windows.
- UI sketches
Sketch screens for an idea before a story exists, and iterate in the conversation until the team shares the same picture.
Settle the decisions before anything is built.
Questions that only surface in the sprint cost the most. Teklens pulls them forward: product, engineering and agents settle them in the same conversation, and you approve the spec.
- Jira story & epic drafting
An idea becomes a story or an epic with acceptance criteria and code context. You review the draft, then Teklens writes it to Jira.
- Decision history & traceability
Decisions and findings are saved as memories on the conversation, artefact, initiative or project. Every result shows which conversation it came from.
- Code owners & effort estimates
Teklens finds who knows an area from git history and estimates effort, complexity and risk of a change against the real code.
- Structured outputs
Estimates, code owners, backlog findings, release notes and review verdicts appear as structured views instead of prose. For dashboards and timelines Teklens generates a view of its own.
- Backlog hygiene
Finds duplicate, stale or contradictory tickets in your Jira and shows what should be merged or closed.
Effort estimate against the real code, with risks and code owner. 
Code owners from git history. 
Backlog findings with a proposal: merge, close, refresh.
Hand over context instead of explaining it again.
The approved spec goes to engineering or a coding agent with its plan and code context. The result comes back as a branch and a pull request, linked in the ticket.
- Coding agent & sandbox
Every agent run works in its own isolated environment. You see progress, model and usage, stop the run or start it again.
- MCP & IDE connections
Connect MCP servers from the public registry to a project so conversations and skills can use more tools. Launch local tools straight from Teklens.
- Test planning
Derives a test plan with edge cases for a story from the code it touches, before implementation starts.
- Implementation review
Checks per acceptance criterion whether a branch or pull request implements the story, with references to the files and symbols involved.
Implementation review per acceptance criterion, with the files involved.
Ship, review and learn from it.
What goes live is documented and checked regularly. Skills and the playbook capture how your team works, so the next round starts better.
- Release notes
Writes release notes or a changelog from what actually shipped in Jira and git since the last release.
- Recurring routines
Ask in a conversation for a recurring task, such as a summary every morning. Each run starts a new conversation with the result.
- Governance & privacy
EU hosting, data separated per tenant, roles for owner, admin and member. Anonymisation per conversation, your own provider keys, approval before every change to Jira.
Release notes from what shipped in Jira and git.
Connected to your tools. Controlled by you.
- Jira
- Confluence
- GitHub
- GitLab
- Watched folders
- MCP servers
More sources are planned. What is not connected, Teklens does not know.
- Hosted in the EU on Cloudflare, data separated per tenant.
- Roles for owner, admin and member per workspace.
- Anonymisation per conversation: personal data is replaced before it reaches the model.
- No logging in the AI gateway. Zero data retention requested where the provider supports it.
- Your own provider keys for OpenAI, Anthropic, Google and xAI.
- Model choice per conversation. Usage visible per person and per workspace.
- Every change to Jira is a draft until you approve it.
Less coordination work.
CPOs and CTOs align AI strategy, operating model and implementation. Product and engineering test the Product Brain with Jira and code. EUR 1,000 in test credits are included.
Two paths: set the direction or test Teklens on real product work.