Amelsis - Indian Contract Intelligence

Amelsis gets its name from Mel, the Latin root for improvement, which is precisely what the software is built to achieve. As a final year law student, I have spent the last six months developing it specifically for the Indian corporate and legal ecosystem. The idea came from watching tier two and tier three law firms in India struggle with enterprise contract tools. They either face massive subscription fees or run into frustrating credit and operation limits, and even then, the software is rarely tailored for Indian contract law.

Traditional legal tech companies pass their heavy server and SaaS infrastructure costs directly onto these smaller firms. Amelsis takes a different approach to keep costs low and privacy intact. Instead of requiring firms to store data on external servers, the software links directly with their existing OneDrive or Google Drive while running locally on their desktops, ensuring everything stays completely off any intermediary cloud server. Contracts are analyzed right where they live, ensuring absolute confidentiality. As it operates, Amelsis continuously learns from the connected OneDrive storage as well as from the specific contracts that are analyzed, adapting its contextual understanding to the firm's drafting style.

The real strength of the platform is its backend. Every prompt and analysis method has been designed alongside practicing corporate advocates who have over five years of experience handling corporate transactions.

To keep things affordable for growing practices, Amelsis operates on a flexible, per-seat license model using a single API. Firms can also use their own API key if they prefer, and we provide hands on technical assistance to help set up and attach the API key directly inside the software. Ultimately, the goal is to make contract lifecycle management accessible and significantly optimize the entire workflow for firms that need it most.

Key features:

1. The eight-stage, 48-item Indian law checklist The only real moat. Everyone can call the same API. What they cannot easily copy is a structured checklist written for Indian statute, with each stage reporting Pass, Partial or Fail and the count of items checked. It turns the output from "the AI thinks this is risky" into "these 48 things were checked and six failed," which is the difference between a toy and something a partner will sign off on.

2. The firm brain Genuinely unique. Running statistics per practice mode on every contract your firm has reviewed, held locally and fed back into the next analysis. Competitors benchmark you against a generic market standard. Amelsis benchmarks you against your own past positions, and the longer a firm uses it the less replaceable it becomes. It is also a switching cost: leaving means leaving the accumulated norms behind.

3. Fully local, bring your own key No Amelsis server exists. Contract text never leaves the firm's machine. For Indian firms handling privileged material this removes the objection that kills most legal AI pilots, and it is architectural rather than a policy promise you have to be trusted on. The flat annual licence with the firm's own API key also means cost is transparent and does not scale with usage, which is the opposite of every per-seat per-document competitor.

4. Anchored editable margin notes The counterparty's Word comments pinned to the exact phrase they mark, with connector lines, sitting alongside Amelsis's own findings colour-graded by risk, all editable in place. Nobody else puts the other side's markup and the AI's markup in the same margin. It means you review in one place instead of toggling between Word and a dashboard.

5. The mode toggle Three lenses that change which checklist items apply and which model is used. It is why the tool does not flag unlimited liability in a promoter SHA or a missing GST clause in an MoU. Most tools apply one generic review to everything and lawyers abandon them after a week of false positives.

6. Annotated Word export Findings come back as native Word comment balloons in the original document. The deliverable is something a lawyer already knows how to use and can forward to a client or the other side without reformatting. Competitors hand you a PDF report that dies in an inbox.

7. Verbatim rendering with track changes Word auto-numbering resolved, tables, footnotes, page geometry, and the other side's tracked changes toggleable between three views. Contract review depends on clause numbers and the cross-references between them, and tools that flatten to plain text break that.

8. Corrected version generation Builds the draft as it would read with every suggestion accepted, and sets it against the original clause by clause. You see the consequence of the advice before taking any of it.